Molecular recognition at the membrane-water interface: Controlling integral peptide helices by off-membrane nucleobase pairing

Article abstract of DOI:10.1021/ja1006349

The aggregation and organization of membrane proteins and transmembrane peptides is related to the interacting molecular species itself and strongly depends on the lipid environment. Because of the complexity and dynamics of these interactions, they are often hardly traceable and nearly impossible to predict. For this reason, peptide model systems are a valuable tool in studying membrane associated processes since they are synthetically accessible and can be readily modified. To control and study the aggregation of peptide transmembrane domains (TMDs) the interacting interfaces of the TMDs themselves can be altered. A second less extensively studied approach targets the TMD assembly by using interaction and recognition of domains at the membrane outside as frequently found in the membrane protein interplay and protein assembly. In the present study, double helical transmembrane domains were designed and synthesized on the basis of a recently reported d,l-alternating peptide pore motif derived from gramicidin A. The highly hydrophobic and aromatic transmembrane peptide was covalently functionalized with a short peptide nucleic acid (PNA) used as specific outer-membrane recognition unit. The PNA sequences were chosen with high polarity to ensure localization within the aqueous phase. To estimate the impact of the membrane adjacent recognition on the TMD assembly by Foerster resonance energy transfer (FRET), fluorescence probes were covalently attached to the side chains of the membrane spanning peptide helices. Dimerization of the TMD-peptide/PNA conjugates within unilamellar lipid vesicles was observed. The dimer/monomer ratio of TMDs can be controlled by temperature variation.

Full text of DOI:10.1021/ja1006349

Published on Web 05/19/2010
Molecular Recognition at the Membrane-Water Interface:
Controlling Integral Peptide Helices by Off-Membrane
Nucleobase Pairing
Philipp Erik Schneggenburger, Stefan Mu¨llar, Brigitte Worbs, Claudia Steinem, and
Ulf Diederichsen*
Institute for Organic and Biomolecular Chemistry, Georg-August-UniVersity Go¨ttingen,
37077 Go¨ttingen, Germany
Received January 24, 2010; E-mail: udieder@gwdg.de
Abstract: The aggregation and organization of membrane proteins and transmembrane peptides is related
to the interacting molecular species itself and strongly depends on the lipid environment. Because of the
complexity and dynamics of these interactions, they are often hardly traceable and nearly impossible to
predict. For this reason, peptide model systems are a valuable tool in studying membrane associated
processes since they are synthetically accessible and can be readily modified. To control and study the
aggregation of peptide transmembrane domains (TMDs) the interacting interfaces of the TMDs themselves
can be altered. A second less extensively studied approach targets the TMD assembly by using interaction
and recognition of domains at the membrane outside as frequently found in the membrane protein interplay
and protein assembly. In the present study, double helical transmembrane domains were designed and
synthesized on the basis of a recently reported D,L-alternating peptide pore motif derived from gramicidin
A. The highly hydrophobic and aromatic transmembrane peptide was covalently functionalized with a short
peptide nucleic acid (PNA) used as specific outer-membrane recognition unit. The PNA sequences were
chosen with high polarity to ensure localization within the aqueous phase. To estimate the impact of the
membrane adjacent recognition on the TMD assembly by Fo¨rster resonance energy transfer (FRET),
fluorescence probes were covalently attached to the side chains of the membrane spanning peptide helices.
Dimerization of the TMD-peptide/PNA conjugates within unilamellar lipid vesicles was observed. The dimer/
monomer ratio of TMDs can be controlled by temperature variation.
Introduction
Nevertheless, this direct approach of subunit interactions via
charges or fit of transmembrane helix topologies does not
Integral membrane proteins represent about 20-30% of all
proteins in sequenced genomes.1 While certain membrane
proteins are functional as monomers, others need to assemble
into oligomeric structures to carry out their biological role. These
membrane protein oligomers form ion channels and pore
structures via arrangement of transmembrane domains and are
often involved in key cellular processes such as protein
translocation and respiration.2 Forces directing and stabilizing
membrane protein oligomerization have been identified in
several experimental studies using model systems of single
spanning transmembrane peptide helices. Sequences like the
Gly-xxx-Gly motif (x ) any other amino acid residue) inducing
a helix topology for proper interaction or single residue
electrostatic recognition promote the ability of individual
transmembrane helix surfaces to adopt close contact sides.3,4
Additionally, interhelical hydrogen bonds between polar amino
acid residues may further stabilize helix-helix contacts.5-11
provide the flexibility many protein assemblies require.12 The
presence of charges in the membrane might impose difficulties
in constructing ion selective pores. In case of the nicotinic
acetylcholine receptor (AChR), the interaction between the
luminal, N-terminal domains is suggested as initial step in the
association of the subunits.12 According to this model, the
recognition of the N-terminal domains and interaction at the
luminal membrane side facilitate the subsequent interaction and
assembly of transmembrane as well as cytoplasmic domains
from different subunits of the AChR machinery.
To our knowledge, very little is reported on model systems
dealing with this specific high-affinity interaction at the
membrane surface providing an organizing and assembling
effect on structured transmembrane domains. Existing model
(6) Walters, R. F.; DeGrado, W. F. Proc. Natl. Acad. Sci. U.S.A. 2006,
103, 13658–13663.
(7) Dawson, J. P.; Weinger, J. S.; Engelman, D. M. J. Mol. Biol. 2002,
316, 799–805.
(1) Wallin, E.; v. Heijne, G. Protein Sci. 1998, 7, 1029–1038.
(2) Rath, A.; Deber, C. M. Proteins 2008, 70, 786–793.
(3) Unterreitmeier, S.; Fuchs, A.; Scha¨ffler, T.; Heym, R. G.; Frishman,
D.; Langosch, D. J. Mol. Biol. 2007, 374, 705–718.
(4) Choma, C.; Gratkowski, H.; Lear, J. D.; DeGrado, W. F. Nat. Struct.
Biol. 2000, 7, 161–166.
(8) North, B.; Christian, L.; Fu Stowell, X.; Lear, J. D.; Saven, J. G.;
DeGrado, W. F. J. Mol. Biol. 2006, 359, 930–939.
(9) Rath, A.; Tulumello, D. V.; Deber, C. M. Biochemistry 2009, 48, 3036–
3045.
(10) Lee, J.; Im, W. J. Am. Chem. Soc. 2008, 130, 6456–6462.
(11) Yano, Y.; Matsuzaki, M. Biochemistry 2006, 45, 3370–3378.
(12) Hall, Z. W. Trends Cell. Biol. 1992, 2, 66–68.
(5) Senes, A.; Engel, D. E.; DeGrado, W. F. Curr. Opin. Struct. Biol.
2004, 14, 465–479.
9
8020 J. AM. CHEM. SOC. 2010, 132, 8020–8028
10.1021/ja1006349  2010 American Chemical Society

Molecular Recognition at the Membrane-Water Interface
A R T I C L E S
limited to dimerization, although it is the simplest case, and it
is, furthermore, suited to evaluate dynamical changes if applied
to interacting transmembrane helices, for example, by adjusting
the directing effect via variation of the respective linker
length.20,21 This would likewise lead to an enhanced understand-
ing why the described process of an initial interaction and
recognition prior to transmembrane helix assembly is favored
in several protein assemblies. Furthermore, the idea of applying
structured PNA/TMD conjugates at the membrane-water
interface is generally applicable to study reversible interaction
processes at membrane surfaces, for example, by modulating
membrane associated substrate conversion pathways that are
mediated by adhesively bound enzymes.
Figure 1. Concept and structural design of outer-membrane PNA nucleo-
base recognition with impact on the aggregation of transmembrane domains
(TMDs), controllable by the applied temperature.17,18
systems either focus on the molecular recognition of species
solely anchored to the membrane surface or on the direct
interaction of transmembrane helices.9,13-16 Since the single
steps of the described recognition and assembling processes are
located in environments of opposing polarities shielded from
each other by the lipid head groups, it is challenging to create
peptide model systems operating at this interface and being
affected by the membrane properties like lipid bilayer composi-
tion, hydrophobic matching, and membrane fluctuation.
A solid phase peptide synthesis approach was selected to
obtain a model system that does not mimic a certain domain of
a native protein species but provides a process analogy to face
general questions concerning the effect of environmental
properties on in-membrane helix assembly that is induced by a
preorganization at the membrane’s adjacent water layers.
Herein, the design and synthesis of an artificial peptide
conjugate is described containing a ꢀ-helical membrane spanning
domain that is structurally derived from the D,L-alternating
gramicidin A pore motif.17 This domain was covalently attached
to a peptide nucleic acid (PNA) recognition moiety by a linker
capable of penetrating the lipid headgroup region. Dimerization
of the TMDs within the lipid environment of unilamellar vesicles
based on temperature dependent PNA recognition at the
membrane outside was studied providing proof-of-concept for
the designed system (Figure 1). The recognition and assembling
process of the transmembrane helices is, thereby, followed
within phospholipid vesicles by means of Fo¨rster resonance
energy transfer (FRET) experiments. With respect to the applied
technique, fluorophores were attached to the N-terminal side
chain of the transmembrane segments.
Experimental Section
The solid phase synthesis of a peptide model system as described
above required the thorough design of the transmembrane domain
with respect to its invariability, geometry concordant with hydro-
phobic matching, and its orientation within the membrane. Single
transmembrane domains were synthesized (Supporting Information)
and studied applying CD and fluorescence spectroscopy. Prior to
synthesis of the entire TMD/PNA constructs, the proper choice of
stability, selectivity, and strand orientation of the PNA double strand
recognition moiety was analyzed by UV spectroscopy determining
oligomer length and sequence. The initial studies were followed
by orthogonal functionalization of the TMD with respective FRET
probes and a PNA recognition unit coupled by polyethylene glycol
(PEG) linker. The donor/acceptor labeled constructs were applied
to the recognition process by incorporation into unilamellar lipid
vesicles and FRET assays.
Preparation of Lipid-Peptide Complexes. Large unilamellar
vesicles (LUVs) of 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC)
were prepared in phosphate buffer (1 mM NaH2PO4/Na2HPO4, pH
) 7.0) following methods of MacDonald et al.28 For CD spectros-
copy, DLPC dissolved in CHCl3 (10 mg/mL) and the peptides
dissolved in MeOH were mixed yielding a solution of CHCl3/MeOH
(1/1, v/v). Concentrations of peptide stocks were determined by
UV absorption. Removing the solvents in a nitrogen stream at
temperatures above the lipid main phase transition temperature of
DLPC (tm ) -2.1 °C)23 produced an almost clear lipid/peptide
film at the test tube walls. After removing of residual solvent under
reduced pressure for 12 h at T > tm, the lipid films were rehydrated
with buffer solution. After 1 h of incubation at T > tm, the hydrated
lipid films were vortexed several times for 30 s with subsequent
incubation for 5 min (5 cycles). The milky suspensions were
extruded 25 times through a polycarbonate membrane (100 nm
nominal pore size) using a miniextruder (Liposofast, Avestin,
Ottawa, Canada) to produce an almost clear vesicle suspension.
Subsequently, the vesicle suspensions were deposited in precision
cells (Quartz Suprasil, Hellma, Mu¨hlheim, Germany). In case of
fluorescence spectroscopy (FRET analysis), DLPC dissolved in
CHCl3 (40 mg/mL) and peptides dissolved in MeOH (concentration
estimated via UV absorption) were mixed under occasional swirling
in the following order: (i) DLPC, (ii) donor fluorophore labeled
species (15, 18), (iii) acceptor fluorophore species (16, 19), (iv)
nonlabeled species (17, 20). The mixtures were heated to 40 °C
prior to solvent removal in the nitrogen stream and treated as
A reliable and reproducible peptide model on basis of the
presented prototype allows addressing the question why nature
uses the mode of protein interaction that was described above
instead of direct in-membrane recognition that has been
extensively studied and was ascribed to a variety of different
driving forces.2,10,19 It can be assumed that the influence, for
example, hydrophobic matching or the directing effect of lipid
domain formation, appears considerably different in case of a
recognition at the membrane’s outside. The use of PNAs is not
(13) Marsden, H. R.; Elbers, N. A.; Bomans, P. H. H.; Sommerdijk,
N. A. J. M.; Kros, A. Angew. Chem. 2009, 121, 2366–2369. Angew.
Chem., Int. Ed. 2009, 48, 2330-2333.
(14) Kashiwada, A.; Matsuda, K.; Mizuno, T.; Tanaka, T. Chem.sEur. J.
2008, 14, 7343–7350.
(15) Stengel, G.; Zahn, R. J. Am. Chem. Soc. 2007, 129, 9584–9585.
(16) Chan, Y.-H. M.; van Lengerich, B.; Boxer, S. G. Proc. Natl. Acad.
Sci. U.S.A. 2009, 106, 979–984.
(20) Krishnan-Ghosh, Y.; Stephens, E.; Balasubramanian, S. J. Am. Chem.
Soc. 2004, 126, 5944–5945.
(21) Wittung, P.; Nielsen, P.; Norde´n, B. J. Am. Chem. Soc. 1997, 119,
3189–3190.
(17) Ku¨sel, A.; Khattari, Z.; Schneggenburger, P. E.; Banerjee, A.; Salditt,
T.; Diederichsen, U. ChemPhysChem 2007, 8, 2336–2343.
(18) Rasmussen, H.; Kastrup, J. S.; Nielsen, J. N.; Nielsen, J. M.; Nielsen,
P. E. Nat. Struct. Biol. 1997, 4, 98–101.
(22) MacDonald, R. C.; MacDonald, R. I.; Menco, B. P. M.; Takeshita,
K.; Subbarao, N. K.; Hu, L. Biochim. Biophys. Acta 1991, 1061, 297–
303.
(19) Tiwari-Woodruff, S. K.; Schulteis, C. T.; Mock, A. F.; Papazian, D. M.
Biophys. J. 1997, 72, 1489–1500.
(23) Oliynyk, V.; Ja¨ger, M.; Heimburg, T.; Buckin, V.; Kaatze, U. Biophys.
Chem. 2008, 134, 168–177.
9
J. AM. CHEM. SOC. VOL. 132, NO. 23, 2010 8021

A R T I C L E S
Schneggenburger et al.
described above. The resulting peptide lipid complexes were kept
at 40 °C until adjusting to room temperature was allowed prior to
extrusion.
ꢁA
F
F0
F
) m
+ (1 - m) 1 - ꢁ
(2)
F0
DꢁD
(
)
( )
m
FRET Analysis. Fluorescence spectra were obtained with a
JASCO FP 5600 fluorescence spectrometer (Gross-Umstadt, Ger-
many) under temperature control using a thermostat (Model 1162A,
VWR International, Darmstadt, Germany). Fluorescence was
excited at 485 nm with a detection of the fluorescence emission
between 495 and 700 nm. The excitation and emission bandwidths
were set to 3 nm, the data pitch was 1 nm, and the response time
was adjusted to 0.2 s. Oligomers 15 and 18 were equipped with
7-nitro-2-1,3-benzoxadiazol-4-yl (NBD) as donor (D) fluorophore
and 5(6)-carboxytetramethylrhodamine (TAMRA) served as ac-
ceptor (A) in oligomers 16 and 19. For temperature dependent
measurements at equimolar concentrations of either peptide/PNA
conjugates 15/16 or the control species 18/19, the total peptide
concentration was adjusted to 5.5 µM. The temperature was varied
between 10 and 80 °C with ∆T ) 10 °C, using a 5 min cycle for
each thermal equilibration. The data shown in Figure 5 were
recorded at a P/L-ratio of 1/1000 in DLPC LUVs. In case of the
concentration dependent FRET-analysis (Figure 6), the concentra-
tion of peptide 15, as well as the total peptide concentration, was
kept constant using the nonlabeled peptide 17. The concentration
of peptide 15 was adjusted to 2.75 µM, while varying the
concentration of 16 from 0.00 to 2.20 µM (DLPC LUVs, P/L )
1/1000). After extrusion of the lipid vesicles, the fluorescence
emission of the sample was measured at 20 °C. After heating and
equilibrating at 60 °C, the sample was measured again, then cooled
and reequilibrated at 20 °C to check for reversibility. For plotting
the donor fluorescence as a function of temperature or the mole
fraction of the acceptor peptide, the fluorescence intensity at 530
nm was displayed. For the control experiment using fluorophore
labeled transmembrane domains 18-20 lacking the PNA recogni-
tion moiety, an identical sample preparation and data treatment was
applied. For the plots of the control, four data sets were averaged
(Supporting Information).
where m is the peptide fraction of the monomeric state and 1 - m
that of the dimeric state. The equation for a monomer-trimer
equilibrium reads (eq 3):
2
ꢁA
ꢁA
F
F0
F
2
) m
+ (1 - m) 1 - ꢁ2D
-
( ) ( )
( )
( )
[
]
F0
ꢁD ꢁD
ꢁD
m
(3)
The dissociation constant KD is defined as (eq 4):
(mꢁP)n
KD )
(4)
(1 - m)ꢁP/n
where ꢁP is the lipid-to-peptide ratio and n the number of species
in the oligomer. KD is expressed in units of peptide/lipid mole
fraction (MF). For the presented simulated plots (Figure 6 C,D),
the Fo¨rster radius was set to R0 ) 5.1 nm as estimated in previous
experiments for the NBD/TAMRA-pair.25
By taking the Fo¨rster radius and the geometry of the peptide
helices with an approximate helix length of 26 Å and a diameter
of 6 Å into account, a fluorescence transfer efficiency of E ) 99.9%
is calculated assuming a parallel assembly of the helices. An
antiparallel peptide assembly would yield E ) 97.9%.26 As in both
cases, for parallel as well as for the antiparallel configuration, the
transfer efficiencies are very high, the two different scenarios cannot
be distinguished. Using this approximation, the statistical occurrence
of FRET in a vesicle without the formation of aggregates was taken
into account (eq 5):
N0
A A
N0
F
) A exp -k ꢁ
R2 + A exp -k ꢁ
R2
(5)
(
)
(
)
1
1
0
2
2
0
A A
(F )
0
m
where A1, A2, k1, k2 are constants, N0 is the total number of peptides,
and A the area of one vesicle.28 The constants are chosen for Re/R0
) 0.28 Re is the radius of closest approach. The area of a DLPC
lipid was assumed to be 0.63 nm2.29
Theoretical Treatment of the FRET Data. The dependence
of the fluorescence emission intensity of donor (NBD)-labeled
peptides on the mole fraction of acceptor (TAMRA)-labeled
peptides in a lipid bilayer enables the determination of the number
of subunits in the peptide assembly.24 This model is based on four
assumptions: (i) labeling does not influence the aggregation of the
peptides, (ii) all peptides are located in the lipid bilayer, (iii) the
interaction of the peptides is random, and (iv) one acceptor is
capable of quenching all donors within one aggregate.25 With a
random number of donors and acceptors in the vesicles, the total
number of quenched (NQ) and nonquenched donors (ND) follows a
binomial distribution. The recorded relative fluorescence (F/F0) is,
thus, related to NQ and ND (eq 1):
Results and Discussion
The design of the construct used for studying the dimerization
of transmembrane helices induced by base pair recognition of
attached outer-membrane oligomers required a conformationally
stable and uniform helix motif. The TMDs were, thus, based
on a recently reported homodimer peptide analogue of the
natural antibiotic gramicidin A (gA).30-32 As indicated by X-ray
crystallography, the peptide H-(YY)4K-OH (1) with D,L-
alternating configuration exhibits an antiparallel, double helical
motif providing a water mediated hydrogen bonding network
of ꢀ-helical tubes.30 In a ꢀ-helix, all peptide side chains are
oriented to the helix outside, and the presented motif provides
heliceswithaperiodicityof5.6residues.ThepeptideH-(FY)5WW-
OH (2) likewise adopts this homodimeric ꢀ5.6-helical structure
fQ NQ
F
F0
) 1 - 1 -
(1)
(
)
fD ND
where F is the fluorescence in the presence of the acceptor labeled
peptides and F0 that in its absence, fD is the molar fluorescence of
nonquenched donor D, and fQ that of quenched donor. With ꢁA +
ꢁD + ꢁu ) 1, where ꢁA denotes the mole fraction of the acceptor
peptide, ꢁD the one of the donor and ꢁu that of the nonlabeled
peptide, and the assumption of a monomer-oligomer equilibrium,
the following expression can be written to describe a monomer-dimer
equilibrium (eq 2):
(26) Naarmann, N.; Bilgic¸er, B.; Meng, H.; Kumar, K.; Steinem, C. Angew.
Chem. 2006, 118, 2650–2653. Angew. Chem., Int. Ed. 2006, 45, 2588-
2591.
(27) You, M.; Li, E.; Wimley, W. C.; Hristova, K. Anal. Biochem. 2005,
340, 154–164.
(28) Wolber, P. K.; Hudson, B. S. Biophys. J. 1979, 28, 197–210.
(29) Kucˇerka, N.; Liu, Y.; Chu, N.; Petrache, H. I.; Tristam-Nagle, S.;
Nagle, J. F. Biophys. J. 2005, 88, 2626–2637.
(30) Alexopoulos, E.; Ku¨sel, A.; Sheldrick, G. M.; Diederichsen, U.; Uso´n,
I. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 1971–1980.
(31) Andersen, O. S. Annu. ReV. Physiol. 1984, 46, 531–548.
(32) Kelkar, D. A.; Chattopadhyay, A. Biochim. Biophys. Acta 2007, 1768,
2011–2025.
(24) Adair, B. D.; Engelman, D. M. Biochemistry 1994, 33, 5539–5544.
(25) Naarmann, N.; Bilgic¸er, B.; Kumar, K.; Steinem, C. Biochemistry 2005,
44, 5188–5195.
9
8022 J. AM. CHEM. SOC. VOL. 132, NO. 23, 2010

Molecular Recognition at the Membrane-Water Interface
A R T I C L E S
and can be incorporated as TMD into lipid bilayers.17 This
membrane active pore motif has been revealed to span the
membrane hydrophobic core with the tryptophan residues
anchoring at the lipid headgroup interfacial region of DLPC
(12:0) and 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC,
14:0) peptide/lipid complexes.17,33
Conformational Restricted Hairpin Linked ꢀ5.6-Helix. As
known from gA, the ꢀ5.6-helix of D/L-alternating peptides exists
in an equilibrium with the ꢀ6.3-form. As the ꢀ5.6-double helical
arrangement spans the entire hydrophobic lipid bilayer core,
the ꢀ6.3-single channel is capable of penetrating one single leaflet
of the membrane while a lateral bilayer dynamic enables a head-
to-head junction assembly that constitutes the channel con-
former. Several functional, linked channel forms (ꢀ6.3-helical)
derived from gA have been reported, for example, using special
amino acid building blocks or tartaric acid residues.34-36 To
preserve the ꢀ5.6-helical pore conformation spanning the entire
hydrophobicbilayercoreandtofurtherexcludeamonomer-dimer
equilibrium, the transmembrane domain H-(FY)5WW-GKPG-
(FY)5WW-OH (3) was introduced. This peptide structure
contains a loop region that covalently links the antiparallel
oriented peptide strands in the double helical conformation. In
this case, the reverse turn with the sequence GKPG emerged to
be the appropriate motif with respect to retaining the ꢀ5.6-helical
pore conformation.37 Flanking tryptophan residues are imple-
mented in order to ensure a membrane spanning orientation via
interfacial anchoring at the glycerol carbonyl sites (beneath the
lipid headgroup) on both sides of the membrane.38 In addition
to positioning of the helix perpendicular to the membrane
surface, tryptophan residues are known to promote a lateral
interaction of membrane spanning peptide helices.39,40 With the
applied design, we have taken advantage of the lateral aggrega-
tion potential that facilitates close contact within the helix
assembly. For this reason, transmembrane helices are designed
to be directed via PNA recognition in the water phase adopting
a side chain mediated parallel orientation within the hydrophobic
bilayer core.
Figure 2. CD spectra of the homodimeric transmembrane domains
H-(FY)5WW-OH (2), H-(FY)5WW-NH2 (4), and Ac-(FY)5WW-NH2 (5)
varying in constitution of the peptide termini, the hairpin transmembrane
domain H-(FY)5WW-GKPG-(FY)5WW-OH (3), and the fluorophore (NBD)
labeled as well as PNA functionalized system H-(K(NBD)
Y(FY)4WW-GK(C2H4O)2-gcgtggKK-H)PG-(FY)5WW-OH (15) (see below)
reconstituted in DLPC vesicles.
(FY)5WW-NH2 (5). A correlation of a lower helical content with
lacking terminal Coulomb interactions was observed. Thus,
charged termini appear to be relevant for intramolecular terminal
interaction between both strands of the nonconnected side of
the hairpin (charge compensation) and for an interaction with
the lipid headgroup interface (polarity). Therefore, the C- and
N-terminus are not addressable for further functionalization of
the TMD with fluorophore labels.
Design of the PNA Recognition System. Aggregation analysis
of the transmembrane domain by means of FRET required
functionalization of the double helical loop structure 3 with
fluorophores NBD or TAMRA. Being aware of conformational
restrictions by attaching the fluorophores to the terminal ends
and the indispensability of free peptide termini, a lysine residue
was introduced as N-terminal amino acid facilitating a side chain
attachment of the fluorophores as amides (Scheme 1). Amino-
ethylglycine (aeg)-PNA was selected as outer-membrane rec-
ognition motif since it is synthetically compatible with Fmoc-
solid phase peptide synthesis (SPPS) that was likewise applied
for the synthesis of the transmembrane domain. Furthermore,
PNA allows duplex formation also with parallel strand orienta-
tion and provides an appropriate duplex stability, selectivity,
and polarity as desired for the model recognition system.43,44
The PNA double strand should be placed in close proximity to
the membrane surface with a reasonable short PNA duplex not
exceeding the TMD length of about 26 Å in order to allow for
application of highly aligned multilamellar membrane stacks
in future experiments.45-47 Attachment of the PNA recognition
unit to the TMD was conducted at the lysine side chain within
the loop region. A diethylene glycol amino acid building block
(Fmoc-HN-(C2H4O)2CH2-COOH) was used to link the water-
soluble PNA oligomer and the highly hydrophobic transmem-
brane domain. This linker is well suited to penetrate the lipid
head groups that separate the regimes of opposing polarity
(Scheme 1). A temperature induced transition from duplex to
single strand PNA was supposed to serve as a reversible trigger
Circular dichroism (CD) analysis of the hairpin structure 3
within large unilamellar DLPC vesicles showed the typical
characteristics of an antiparallel ꢀ5.6-helical double strand with
maxima at 228 and 205 nm and a negative Cotton effect below
200 nm.41,42 The decrease in molar ellipticity compared to
homodimer 2 can be assigned to missing charges due to the
reverse turn modification substituting the peptide termini that
are normally charged under physiological conditions (Figure
2).33 This hypothesis was validated by CD spectra comparison
of oligomer 2 with terminal noncharged N-acyl and/or C-amide
functionalized homodimers H-(FY)5WW-NH2 (4) and Ac-
(33) Schneggenburger, P. E.; Beerlink, A.; Worbs, B.; Salditt, T.; Dieder-
ichsen, U. ChemPhysChem 2009, 10, 1567–1576.
(34) Essen, L.-O.; Koert, U. Annu. Rep. Prog. Chem., Sect. C: Phys. Chem.
2008, 104, 165–188.
(35) Xie, X.; Al-Momani, L.; Reiꢀ, P.; Griesinger, C.; Koert, U. FEBS J.
2005, 272, 975–986.
(36) Stankovic, C. J.; Heinemann, S. H.; Schreiber, S. L. J. Am. Chem.
Soc. 1990, 112, 3702–3704.
(37) Sastry, M.; Brown, C.; Wagner, G.; Clark, T. D J. Am. Chem. Soc.
2006, 128, 10650–10651.
(38) De Planque, M. R. R.; Killian, J. A. Mol. Membr. Biol. 2003, 20,
271–284.
(43) Nielsen, P. E.; Egholm, M.; Berg, R. H.; Buchardt, O. Science 1991,
254, 1497–1500.
(39) Ridder, A.; Skupjen, P.; Unterreitmeier, S.; Langosch, D. J. Mol. Biol.
2005, 354, 894–902.
(44) Nielsen, P. E.; Egholm, M. Peptide Nucleic Acids-Protocols and
Applications, Horizon Scientific Press, Wymondham, 1999.
(45) Salditt, T. Curr. Opin. Struct. Biol. 2003, 13, 467–478.
(46) Mennicke, U.; Salditt, T. Langmuir 2002, 18, 8172–8177.
(47) Fahsel, S.; Pospiech, E.-M.; Zein, M.; Hazlet, T. L.; Gratton, E.;
Winter, R. Biophys. J. 2002, 83, 334–344.
(40) Adamian, L.; Liang, J. J. Mol. Biol. 2001, 311, 891–907.
(41) Koeppe II, R. E.; Andersen, O. S. Annu. ReV. Biophys. Biomol. Struct.
1996, 25, 231–258.
(42) Chen, Y.; Wallace, B. A. Eur. Biophys. J. 1997, 26, 299–306.
9
J. AM. CHEM. SOC. VOL. 132, NO. 23, 2010 8023

A R T I C L E S
Schneggenburger et al.
Scheme 1. Exemplarily SPPS of TMD/PNA Conjugate 15 with NBD-Donor Labeling
for the assembly and disassembly of the transmembrane
domains. Thus, the stability of the PNA double strand, that is,
its melting temperature Tm, was chosen to be settled well above
the lipid main phase transition temperature tm to ensure that
the bilayers composed of DLPC (tm ) -2.1 °C)23 or DMPC
(tm ) 23.6 °C)48 are continuously adopting the liquid LR-phase
unaffected by the temperature dependent pairing and melting
of the PNA double strand. Summarizing, the PNA moiety should
be rather short in sequence regarding its applicability for
different substrates, further being guanine-cytosine rich to reach
the aspired melting transition interval of 30-55 °C. Addition-
ally, the PNA strands should contain a sufficient number of base
pairs to still enable a selective recognition process avoiding
higher ordered aggregation like guanine tetrad formation.49
Several short PNA sequences lacking the TMD were syn-
thesized in order to evaluate the differences in stability between
parallel and antiparallel strand orientation of respective double
strands and to further estimate the appropriate length with
respect to requirements mentioned before. PNA octamers
H-(Gly)3-cgcttacc-(Lys)2-NH2 (6) and H-(Gly)3-ggtaagcg-(Lys)2-
NH2 (7) formed an antiparallel double strand (Tm ) 72.4 °C),
whereas parallel strand orientation (Tm ) 72.7 °C) was provided
with PNA 6 and H-(Gly)3-gcgaatgg-(Lys)2-NH2 (8). The stabili-
ties of PNA octamers were too high to trigger temperature
dependent disassembly. Preferentially, PNA aggregates with
parallel strand orientation were investigated having both TMDs
(48) Heimburg, T. Thermal Biophysics of Membranes, Wiley-VCH, Wein-
(49) Datta, B.; Bier, M. E.; Roy, S.; Armitage, B. A. J. Am. Chem. Soc.
2005, 127, 4199–4207.
heim, 2007.
9
8024 J. AM. CHEM. SOC. VOL. 132, NO. 23, 2010

Molecular Recognition at the Membrane-Water Interface
A R T I C L E S
aeg-G(Bhoc)-monomer; for this building block, side reactions
were most prominent. Self-capping by ‘acetyl-nucleobase’
species during the neutralization/deprotection step was avoided
by NMP/DIPEA washing introduced after the capping cycle.51
At the N-terminal end of the PNA sequence, two lysine
residues were attached to give construct 14. They were
introduced for enhancement of the segment polarity and should
ensure localization within the membrane adjacent water layers.
Finally, the fluorophore labeled constructs H-(K(NBD)Y(FY)4-
WW-GK(C2H4O)2-gcgtggKK-H)PG-(FY)5WW-OH (15) and
H-(K(TAMRA)Y(FY)4WW-GK(C2H4O)2-cgcaccKK-H)PG-
(FY)5WW-OH (16) were obtained being complementary in their
PNA nucleobase sequences. Purification was performed via
dialysis in combination with RP-HPLC using a C4 stationary
phase. Structural integrity of the oligomers was proven by high-
resolution mass spectrometry. Details regarding synthesis,
purification, and characterization are given in the Supporting
Information.
Figure 3. Melting curve obtained by UV spectroscopy for the equimolar
mixture of PNA hexamers 9 and 10 (10 mM NaH2PO4/Na2HPO4, 100 mM
NaCl, pH ) 7.0) used as recognition unit with Tm ) 41 °C.
aligned and attached at the respective C-terminal end. UV
analyses of the duplex formed by parallel oriented hexamers
H-(Gly)3-cgcacc-(Lys)2-NH2 (9) and H-(Gly)3-gcgtgg-(Lys)2-
NH2 (10) revealed a melting temperature of 41 °C (Figure 3).
This stability is in the appropriate range for serving as
recognition device within the envisioned model system. This
Tm is high enough to ensure standard phospholipids to remain
in the fluid membrane state, and is still applicable as trigger
parameter. For the melting experiments, the PNA sequences
were investigated without the TMD attached, since the TMD/
PNA constructs were insufficiently soluble in aqueous media
requiring the lipid environment for solubilization of the
hydrophobic transmembrane segment.
Synthesis of TMD/PNA Constructs. Preparation of the resin
bound, fully protected transmembrane hairpin was provided by
automated microwave-assisted Fmoc-SPPS. Double coupling
was performed for the tryptophan residues. Capping of nonre-
acted free amine functionalities was only performed until
incorporation of the Mtt-protected lysine residue in order to
avoid acylation of the secondary amine functionality.50 Lysine
side chains were orthogonally protected to enable attachment
of the fluorescence probes at the side chain of the N-terminal
lysine residue as well as PNA functionalization at the lysine
residue of the turn region in the following step. Therefore, the
Fmoc- and the 4-methyltrityl (Mtt)-protection groups were
applied during the assembly of the transmembrane domain
allowing regioselective diversification (Scheme 1). As protecting
group pattern of the terminal lysine residue, Fmoc-protection
at the side chain and Boc-protection at the N-terminus were
used. Accomplishing the automated synthesis of resin bound
TMD 11 followed by Fmoc-deprotection, the FRET probe
4-fluoro-7-nitrobenzofurazan (NBD-F) was coupled to the side
chain amino functionality of the N-terminal lysine residue
yielding peptide 12 (Scheme 1). This was followed by selective
deprotection of the highly acid labile lysine Mtt-group in the
GKPG-turn, at which the diethylene glycol amino acid was
subsequently attached. For elongation of the resin bound TMD
domain 13 with the PNA sequence via manual SPPS (Fmoc/
Bhoc-strategy), some precautions had to be taken according to
deprotection and capping cycles. To suppress base-catalyzed
transamidation rearrangement by the nascent primary amine that
would either lead to base deletion or sequence termination,42
the deprotection times were kept short and double couplings
with diminished coupling times were performed for the Fmoc-
In addition to the PNA sequence complementary constructs
15 and 16 carrying the donor and acceptor fluorescence probe,
respectively, a third construct 17 with identical PNA sequence
compared to the acceptor-species 16 but lacking the fluorescence
label was synthesized. The nonlabeled species 17 was acylated
at the side chain position of the terminal lysine of the
transmembrane domain (Figure 4). It was used to ensure a
constant total peptide concentration and, therefore, a constant
peptide-to-lipid ratio throughout the mixing experiments within
the lipid bilayer.
The ꢀ5.6-helical TMD structure of peptide 15 was exemplarily
verified by CD spectroscopy after incorporation in DLPC
vesicles (Figure 2). The CD spectrum still revealed high ꢀ5.6-
helix propensity besides contributions from the PNA strand.
TMD reconstitution of oligomers 15 and 16 in lipid bilayers
with a bilayer spanning orientation was proven via analysis of
intrinsic tryptophan fluorescence (Supporting Information) and
more importantly, in-house as well as anomalous X-ray scat-
tering experiments.52,53 The tryptophan fluorescence was moni-
tored for peptides inserted into LUVs composed of DLPC at a
peptide-to-lipid (P/L) ratio of 1/600. A hypsochromic shift of
the tryptophan fluorescence relative to that of tryptophan in
aqueous solution (λmax,em ) 350 nm) indicated the tryptophan
residues to be located in a more hydrophobic environment.54,55
The maximum of the fluorescence emission band was located
at 336 nm for peptide 15 and at 338 nm in case of the equimolar
mixture of peptides 15 and 16. This is in line with the preference
of tryptophan to be located at the bilayer-water interface.56-60
(51) Koch, T.; Hansen, H. F.; Andersen, P.; Larsen, T.; Batz, H. G.; Otteson,
K.; Økrum, H. J. Pept. Res. 1997, 80–88.
(52) Khattari, Z.; Brotons, G.; Arbely, E.; Arkin, I. T.; Metzger, T. H.;
Salditt, T. Physica B 2006, 86, 1521–1531.
(53) Schneggenburger, P. E. Unpublished results, 2010.
(54) Ren, J.; Lew, S.; Wang, J.; London, E. Biochemistry 1999, 38, 5905–
5912.
(55) Lew, S.; Ren, J.; London, E. Biochemistry 2000, 39, 9632–9640.
(56) Chattopadhyay, A.; Raghuraman, H. Curr. Sci. 2004, 87, 175–180.
(57) Bong, D. T.; Janshoff, A.; Steinem, C.; Ghadiri, M. R. Biophys. J.
2000, 78, 839–845.
(58) Strandberg, E.; Morein, S.; Rijkers, D. T. S.; Van der Wel, P. C. A.;
Liskamp, R. M. J.; Killian, J. A. Biochemistry 2002, 41, 7190–7198.
(59) De Planque, M. R. R.; Boots, J.-W. P.; Rijkers, D. T. S.; Liskamp,
R. M. J.; Greathouse, D. V.; Killian, J. A. Biochemistry 2002, 41,
8396–8404.
(60) De Planque, M. R. R.; Demmers, J. A. A.; Bonev, B. B.; Koeppe II,
R. E.; Greathouse, D. V.; Separovic, F.; Watts, A.; Killian, J. A.
Biochemistry 2003, 42, 5341–5348.
(50) Stafforst, T.; Diederichsen, U. Eur. J. Org. Chem. 2007, 681–688.
9
J. AM. CHEM. SOC. VOL. 132, NO. 23, 2010 8025

A R T I C L E S
Schneggenburger et al.
Figure 4. Molecular representation of the acceptor labeled and nonlabeled constructs.
Figure 5. Temperature dependent FRET experiment in DLPC-LUVs. (A) The NBD-fluorescence emission at 538 nm is displayed as a function of temperature
for the equimolar mixture of the control species 18 (NBD)/19 (TAMRA) lacking the PNA recognition moieties (black data points) and the equimolar
mixture of the PNA/TMD constructs 15 (NBD)/16 (TAMRA) (red data points). The experiments were performed at a total peptide concentration of 5.5 µM
and a P/L-ratio of 1/600. (B) Ratio of the donor fluorescence of peptides 15/16 and peptides 18/19 as a function of temperature.
Temperature Controlled Recognition of Peptide Helices in
Lipid Vesicles. Besides analytical ultracentrifugation,4,61-63
chromatographic,4 and spectroscopic methods,62 FRET experi-
ments have been successfully applied to evaluate the oligomeric
state of peptides not only in a micellar environment,63 but also
within lipid bilayers.25-27,64 For instance, several variants of
the GCN4 (leucine zipper) motif have been independently
studied by the above-mentioned methods giving essentially the
same results.25,26,61-63 Since dynamic interactions of an entire
molecular ensemble are in the scope of the presented study
focusing on switching between two oligomerization states, a
detergent-free setup was favored.24-28 Therefore, we made use
of FRET between the donor (NBD) and acceptor (TAMRA)
labeled TMD/PNA constructs 15 and 16 to elucidate aggregate
formation of the TMD peptides within large unilamellar DLPC
vesicles by molecular recognition of the PNAs outside the
membrane.
(FY)5WW-OH (19), both lacking the PNA recognition unit, were
inserted into DLPC vesicles at an equimolar ratio and the donor
fluorescence intensity was recorded at 538 nm. The donor
fluorescence intensity decays with increasing temperature
(Figure 5A), which is attributed to an increase in statistical
quenching at elevated temperature and by a decrease in the NBD
quantum yield with increasing temperature.65 If, however, the
PNAs containing TMD peptides 15 and 16 are present in the
vesicles, the donor fluorescence intensity is considerably smaller
at low temperature (Figure 5A) than for the PNA lacking TMD
peptides 18 and 19. The low donor fluorescence can be
explained by an efficient energy transfer to the acceptor. For
vesicles containing equimolar concentrations of 15 and 16, the
decrease in donor fluorescence upon increasing the temperature
is less pronounced and shows a change in slope between 40
and 50 °C, which is more easily observed in Figure 5B, where
FNBD(15/16)/FNBD(18/19) is plotted. This change in slope is a result
of a partial recovery of the donor fluorescence at elevated
temperature consistent with the notion that the donor-acceptor
distance increases. Comparing the temperature dependence with
the melting curve of pure PNAs 9/10 (see Figure 3), this
observation can be assigned to a melting of the PNA double
strands followed by a dissociation of the TMD peptides. A
contribution from a lipid phase transition can be ruled out (tm
(DLPC) ) -2.1 °C).
It is expected that a change in temperature affects the
aggregation of the TMD peptides as a result of pairing or melting
of the PNA double strands. To prove this hypothesis, temper-
ature dependent FRET-experiments were performed. First, the
fluorophore-labeled species (H-(K(NBD)Y(FY)4WW-GK(Ac)PG-
(FY)5WW-OH (18) and H-(K(TAMRA)Y(FY)4WW-GK(Ac)PG-
(61) Gratkowski, H.; Lear, J. D.; DeGrado, W. F. Proc. Natl. Acad. Sci.
U.S.A. 2001, 98, 880–885.
Base pair recognition of the PNA strands is supposed to result
in dimeric TMD/PNA assemblies. The oligomerization state can
be assessed by variation of the acceptor-peptide fraction (ꢁA)26
at a constant total peptide concentration (Figure 6). Compound
17 with the acylated lysine side chain instead of the attached
fluorescence label was used to replace the TAMRA-labeled
(62) Gratkowski, H.; Dai, Q.-h.; Wand, A. J.; DeGrado, W. F.; Lear, J. D.
Biophys. J. 2002, 83, 1613–1649.
(63) Lear, J. D.; Stouffer, A. L.; Gratkowski, H.; Nanda, V.; DeGrado,
W. F. Biophys. J. 2004, 87, 3421–3429.
(64) Merzlyakov, M.; You, M.; Li, E.; Hristova, K. J. Mol. Biol. 2006,
358, 1–7.
(65) Fleming, K. G. J. Mol. Biol. 2002, 323, 563–571.
9
8026 J. AM. CHEM. SOC. VOL. 132, NO. 23, 2010

Molecular Recognition at the Membrane-Water Interface
A R T I C L E S
Figure 6. (A and B) Fluorescence emission spectra of the NBD-donor labeled compound 15 (2.75 µM) with a maximum at 537 nm and varying amounts
of the acceptor species 16 (from ꢁA ) 0.00 to ꢁA ) 0.40) with a maximum at 571 nm recorded at 20 °C (A) and 60 °C (B). The nonlabeled compound 17
was added at various amounts to yield a total peptide concentration of 5.5 µM and a P/L ratio of 1/1000. (C and D) Plots of the relative changes in
NBD-fluorescence emission as a function of ꢁA for the TMD/PNA (15/16/17) and TMD control peptides (18/19/20) at (C) 20 °C and (D) 60 °C. The black
lines are calculated plots based on the assumption of peptide monomers (dotted line), dimers, trimers (solid lines) or a monomer-dimer equilibrium (dashed
line) with a dissociation constant given in mole fractions (MF).
peptide 16, thus, keeping the P/L-ratio constant. The fluores-
cence emission spectra at T < Tm (T ) 20 °C) show a decrease
in donor fluorescence, while the acceptor fluorescence increases
with higher concentration of the acceptor species 16 (Figure
6A) indicating the occurrence of a FRET process. The ratio
between maximum donor and acceptor fluorescence is different
at 20 and 60 °C (Figure 6A,B), which might be a result of the
different temperature dependencies of the NBD and TAMRA
extinction coefficients. However, this is canceled out as only
relative changes of the fluorescence intensity are plotted. A plot
of the relative fluorescence F/F0 (ꢁA) of the NBD-labeled donor
peptides, with F0 being the fluorescence of the acceptor-free
sample, shows an increased quenching of the NBD-fluorescence
with increasing acceptor concentration. To evaluate the data and
elucidate which aggregates might have been formed as a
function of temperature, several scenarios are discussed: (i)
Assuming that only peptide monomers are present in the
vesicles, statistical occurrence of FRET due to the presence of
several acceptors in any given vesicle has to be taken into
consideration. With a Fo¨rster radius of R0 ) 5.1 nm (see
theoretical treatment of the FRET data),26 the plots shown in
Figure 6C,D (black dotted lines, monomer) are calculated. For
peptides lacking the PNA sequences (18, 19, and 20: H-
(K(Ac)Y(FY)4WW-GK(Ac)PG-(FY)5WW-OH), the calculated
plot is slightly above the data points obtained at 20 °C (Figure
6C), which is probably just a result of the variance of the data
points. The calculated plot is in very good agreement with the
data points at 60 °C (Figure 6D). This demonstrates that all
peptides are in the monomeric state at 60 °C, while at 20 °C,
one can state that the majority is monomeric. (ii) Assuming a
binomial distribution of donors and acceptors in a dimeric
assembly as a result of base pair recognition results in the plots
shown in Figure 6C,D with a significantly larger slope (black
solid lines, dimer). The calculated curve is in rather good
accordance with the data points obtained for the TMD/PNA
construct (15/16/17) at 20 °C but does not fit the data well at
60 °C. This lets us assume that the TMD/PNA construct is
almost completely assembled in a dimeric fashion at 20 °C,
while it is neither purely monomeric nor dimeric at 60 °C. (iii)
It is more likely that instead of stable dimers, a monomer-dimer
equilibrium of the TMD/PNA construct is established at 60 °C.
A curve assuming a monomer-dimer equilibrium with a
dissociation constant of 10-3 MF was calculated, shown in
Figure 6D (black dashed line), which follows the same trend
as the data points.28,26 (iv) To rule out that even larger aggregates
are formed, we also calculated plots of F/F0 (ꢁA) for the
formation of trimers (Figure 6C,D, black solid lines, trimer). It
is obvious that a considerable larger change of F/F0 is expected,
which could not be found in the experiments.
From the overall results, we conclude that the degree of
association/dimer formation is a function of base pair recognition
in the aqueous phase, which is temperature dependent. While
at 20 °C, the data points reveal that a majority of TMD/PNA
peptides are in the dimeric state, clear monomer-dimer equi-
librium is established at 60 °C. Apparently, a complete
dissociation of the dimer assembly is not achieved even at a
temperature well above the melting temperature of the PNA
double strand 9/10 lacking the transmembrane domain. It is very
likely that the lipid bilayer itself influences the TMD association.
Furthermore, the transmembrane helix aggregation might con-
9
J. AM. CHEM. SOC. VOL. 132, NO. 23, 2010 8027

A R T I C L E S
Schneggenburger et al.
tribute to the overall stability, since aggregation was also
observed during chromatographic purification of various TMD
constructs.
respect to a direct interaction within the hydrophobic membrane
core.3,10 Since cellular key functions are driven by this poorly
understood mechanism, the application of the novel peptide
model and its variants will enable a learning process about the
dynamics, organization, and geometrical preferences of mem-
brane bound species and, therefore, provide insight into the
biophysical demands that have to be fulfilled to exhibit
biochemical functioning.
Conclusion
The rational design of a peptide motif providing a transmem-
brane part and a PNA unit for specific recognition was
introduced to study helix-helix interactions within lipid bilayer
structures at the membrane-water interface. Different constructs
with respect to labeling and recognition sequences were
assembled by means of SPPS in combination with orthogonal
fluorophore and linker functionalization. A peptide hairpin with
a stable ꢀ5.6-double helical conformation derived from a D,L-
alternating gramicidin A analogue served as membrane spanning
domain. In combination with the attached PNA recognition unit,
dimerization of these TMD/PNA constructs was observed in
large unilamellar DLPC-vesicles. Monitoring of dimer-monomer
equilibrium was feasible by temperature and concentration
dependent FRET analysis of NBD and TAMRA labeled species
that revealed a dimeric assembly at 20 °C and a fractional
depairing at elevated temperature providing a smaller dissocia-
tion constant.
The novel TMD/PNA conjugate composes the first synthesis
of a membrane associated peptide model system that uses
distinct recognition at the membrane outside to study the in-
membrane dimer-formation of structured domains that are
spanning the entire lipid bilayer core. With such a model, it is
possible to investigate the dependency of the above-described,
native recognition/assembly process from various, especially
lipid bilayer related parameters. These are, for example,
hydrophobic matching, interfacial anchoring, and lipid domain
formation. Up to now, such properties were merely studied with
The respective constructs, possibly equipped with modified
recognition units, will allow for more extensive, reliable
dimer-monomer switching. The potential fusogenicity of the
constructs and expansion of the outer-membrane recognition
unit leading to higher organized transmembrane domains such
as pore assemblies is also in the scope of future work. This
could be achieved by promoting the formation of PNA triple
helices by sequence selection or by applying guanine tetrad
assemblies.20,21,49
Acknowledgment. Generous support of the Deutsche Fors-
chungsgemeinschaft (SFB 803) is gratefully acknowledged. P.E.S.
is grateful for a doctoral bridging stipend of the Go¨ttingen Graduate
School for Neuroscience and Molecular Biosciences (GGNB). We
dedicate this paper to Professor Horst Kessler on the occasion of
his 70th birthday.
Supporting Information Available: Materials and methods,
peptide/PNA synthesis and purification, notes on purification,
analytical data, CD and UV spectroscopic methods, fluorescence
spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA1006349
9
8028 J. AM. CHEM. SOC. VOL. 132, NO. 23, 2010