Báo cáo khoa học: Structure and positioning comparison of two variants of penetratin in two different membrane mimicking systems by NMR pdf

Structure and positioning comparison of two variants of penetratin
in two different membrane mimicking systems by NMR
Mattias Lindberg, Henrik Biversta
˚
hl, Astrid Gra¨ slund and Lena Ma¨ ler
Department of Biochemistry and Biophysics, The Arrhenius Laboratories, Stockholm University, Sweden
The Antennapedia homeodomain protein of Drosophila has
the ability to penetrate biological membranes and the third
helix of this protein, residues 43–58, known as penetratin
(RQIKIWFQNRRMKWKK-amide) has the same trans-
locating properties as the entire protein. The variant, RQI
KIFFQNRRMKFKK-amide, here called penetratin
(W48F,W56F) does not have the same ability. We have
determined a solution structure of penetratin and investi-
gated the position of both peptides in negatively charged
bicelles. A helical structure is seen for residues Lys46 through
Met54. The secondary structure of the variant penetra-
tin(W48F,W56F) in bicelles appears to be very similar.
Paramagnetic spin-label studies and analysis of NOEs
between penetratin and the phospholipids show that pene-
tratin is located within the bicelle surface. Penetratin
(W48F,W56F) is also located inside the phospholipid bicelle,
however, with its N-terminus more deeply inserted than that
of wild-type penetratin. The subtle differences in the way the
two peptides interact with a membrane in an equilibrium
situation could be important for their translocating ability.
As a comparison we have also investigated the secondary
structure of penetratin(W48F,W56F) in SDS micelles and
the results show that the structure is very similar in SDS and
bicelles. In contrast, penetratin(W48F,W56F) and penetra-
tin appear to be located differently in SDS micelles. This

clearly shows the importance of using realistic membrane
mimetics for investigating peptide–membrane interactions.
Keywords: cell-penetrating peptide; penetratin; pAnt;
NMR; bicelle.
Cell-penetrating peptides, CPPs, have the ability to trans-
locate various cell membranes with high efficiency. If they
are covalently linked to a ÔcargoÕ, they still retain their
translocating properties, making them suitable for trans-
porting large cargoes, such as polypeptides and oligonucleo-
tides, across membrane barriers. These properties have
made CPPs interesting for use as vectors for delivery of
hydrophilic biomolecules and drugs into the cytoplasmic
and nuclear compartments of the cell, both in vivo and
in vitro [1–3].
The 60 amino acid residue DNA-binding domain of the
Drosophila transcription factor translocates membranes.
The peptide corresponding to the residues of the third helix
of the Antennapedia homeodomain of Drosophila (residues
Arg43 through Lys58: RQIKIWFQNRRMKWKK) has
been shown to have the same translocating properties as the
entire protein [4–7]. The peptide, known as penetratin, has
the ability to carry large cargoes, such as oligonucleotides,
proteins or other peptides, through biological membranes
[1]. Penetratin is a well-studied CPP, both with regards to its
translocating properties [5–8] and its induced secondary
structure in various membrane mimetic solvents, such as
detergent micelles and phospholipid vesicles [9–13]. The
translocation process does not seem to require a chiral
receptor and the detailed mechanism is still not understood.
Knowledge about the interaction between the peptide and

the membrane is fundamental for the understanding of
the translocation process. Therefore studies in a realistic
membrane environment are important.
A study of the secondary structure of penetratin in a
membrane-like environment with negatively charged SDS
micelles has previously been conducted [14] where it was
shown that penetratin interacts with the SDS micelle and
adopts an a-helical structure in the micelle environment. In
a positioning study using paramagnetic probes it was shown
that penetratin is located with its most N-terminal residues
at the micellar surface and with the C-terminus hidden
inside the interior of the micelle [10]. There is evidence that
the induced secondary structure of a transport peptide is not
an important factor for the transport ability [6,15], but very
little is known about the importance of the positioning
in the membrane. When changing the two tryptophans
(residue 48 and 56) of penetratin to phenylalanines it was
shown that the translocating property of penetratin was
essentially lost [13].
In this study, we have determined a NMR solution
structure of penetratin and investigated the position of
penetratin in negatively charged phospholipid bicelles. The
secondary structure of the nontranslocating analog, denoted
penetratin(W48F,W56F), in negatively charged bicelles has
also been investigated together with the position of the
peptide relative to the surface and interior of the bicelles.
Correpsondence to L. Ma
¨
ler, Department of Biochemistry and
Biophysics, The Arrhenius Laboratories for Natural Sciences,

Stockholm University, SE-106 91 Stockholm, Sweden.
Fax: + 46 (0)8155597, E-mail:
Abbreviations: CPP, cell-penetrating peptide; DHPC, 1,2-dihexanoyl-
sn-glycero-3-phosphatidylcholine; DMPC, 1,2-dimyristoyl-
sn-glycero-3-phosphatidylcholine; DMPG, 1,2-dimyristoyl-sn-
glycero-3-phospho-1-glycerol; DMPS, dimyristoyl-sn-glycero-
3-phosphatidylserine; TSPA, 3-trimethylsilyl-propionic acid-d
4
.
Note: a web site is available at
(Received 3 March 2003, revised 11 April 2003, accepted 23 May 2003)
Eur. J. Biochem. 270, 3055–3063 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03685.x
Bicelles are formed by mixing two phospholipid compo-
nents with different chain lengths (e.g. DHPC, dihexanoyl-
sn-glycero-phosphocholine and DMPC, dimyristoyl-
sn-glycero-phosphocholine) and it has been established that
a mixture of DMPC and DHPC produces disk-shaped
bicelles [16–18]. The size of the bicelle can be controlled by
varying the lipid composition (q ¼ [DMPC]/[DHPC]) and
a fraction of the DMPC can be replaced by charged lipids
[19]. Smaller isotropic bicelles have been shown to retain
their disk-like shape [20,21] properties and are suitable for
high-resolution NMR work in which peptides that associate
with membranes can be studied [22–26]. Here we present an
NMR study of penetratin and the nontranslocating analog
in a q ¼ 0.5 bicellar solution with a fraction of the DMPC
replaced by the negatively charged phospholipid DMPG.
Because penetratin has previously been studied in SDS
micelles we also investigated the structure and position of
penetratin(W48F,W56F) in SDS. The SDS micelle may be

considered as a simple mimic of the amphiphilic environ-
ment of a phospholipid bilayer and its dimensions are
comparable with those of the peptide, which may influence
the interaction between the two and the resulting complex.
The phospholipid bicelles are more membrane-like than a
SDS micelle and may therefore be a better system for
positioning studies of membrane bound peptides. We have
compared structure and positioning results obtained using
the two types of solvents, micelles and bicelles. Based on our
present knowledge we conclude that although secondary
structure induction is quite similar in the two solvents, the
positioning experiments give a more coherent picture in the
bicellar solvent.
Experimental procedures
Sample preparation
Penetratin and penetratin(W48F,W56F) were obtained as
HPLC-purified custom syntheses from Neosystem Inc. and
were used without further purification. Deuterated SDS
was purchased from Cambridge Isotopes Laboratories
Inc. Deuterated phospholipids, dihexanoyl-sn-glycero-
3-phosphatidylcholine-d
22
(DHPC), 1,2-dimyristoyl-sn-
glycero-3-phosphatidylcholine-d
54
(DMPC), 1,2-dimyris-
toyl-sn-glycero-3-phospho-1-glycerol-d
54
(DMPG), 1,2-
dimyristoyl-sn-glycero-3-phosphatidylserine-d

54
(DMPS)
and the spin-labeled phospholipids 1-palmitoyl-2-stearoyl-
sn-glycero-5-doxyl-3-phosphatidylcholine and 1-palmitoyl-
2-stearoyl-sn-glycero-12-doxyl-3-phosphatidylcholine were
purchased from Avanti Lipids. The 5- and 12-doxyl stearic
acids were obtained from Sigma and the MnCl
2
from
Merck.
Bicelle samples were produced by mixing a 0.96
M
aqueous solution of DHPC with a slurry of the long-
chained lipids (DMPC and DMPG or DMPS) in H
2
Oto
obtain a sample with a total lipid concentration of 15%
(w/v) and q ¼ 0.5, which indicates a bicelle diameter of
80–100 A
˚
[16]. Penetratin was added together with the long-
chained lipids to reach a peptide concentration of 3 m
M
.
The pH was checked and adjusted to around 5.5 for each
sample and finally, aqueous KCl was added to a final salt
concentration of 50 m
M
. In experiments using paramagnetic
probes either small amounts of aqueous solution of MnCl

2
,
or small amounts of 5-doxyl- or 12-doxyl-labeled phos-
pholipid in methanol-d
4
wasaddedtothesample.Inall
NMR samples, 25 lLD
2
O was added for field/frequency
locking.
The peptide/SDS samples were prepared by dissolving
the peptide powder at 2 m
M
concentration in 300 m
M
deuterated SDS solution in H
2
O/D
2
O. Under the conditions
used, SDS forms stable micelles with an approximate
number of 60 SDS molecules per micelle [27]. The H
2
O/
D
2
O-ratio was 90 : 10 and the pH was set to 4.1 by adding
small amounts of HCl. The samples used in spin-label
experiments were prepared by adding either small amounts
of aqueous MnCl

2
, or small amounts of 5-doxyl or
12-doxyl-labeled stearic acid dissolved in methanol-d
4
.
Spectroscopy
All NMR experiments were performed at 37 or 45 °Con
Varian Unity spectrometers operating at 600 MHz
1
H
frequency. The chemical shifts were referenced to internal
3-(trimethylsilyl)-propionic acid-2,2,3,3-d
4
(TSPA). Two-
dimensional phase-sensitive spectra were collected using the
method by States and coworkers [28]. The spectral widths in
the two-dimensional experiments were 9000 Hz in both
dimensions and typically the number of complex points
collected in the x
2
dimension was 4096 and 512 in the x
1
dimension. The data were zero-filled to 8192 points in the x
2
dimension and to 4096 points in the x
1
dimension prior to
Fourier transformation. NOESY experiments [29] were
recorded with mixing times of 100, 150 and 300 ms and
TOCSY experiments [30] were recorded with mixing times

of 30, 60 and 90 ms. Water-suppression was achieved with
low-power presaturation or with the WATERGATE
sequence [31]. The data were processed and analyzed using
the
VNMR
program on a Sun sparc5 work station and
with the
FELIX
software (version 2000, Accelrys Inc.). The
chemical shift assignments for penetratin in bicelles and for
penetratin(W48F,W56F) in bicelles and SDS have been
deposited with the BioMagResBank under accession num-
bers 5542 and 5543, respectively.
CD measurements were made on a Jasco J-720 CD
spectropolarimeter using a 0.01-mm quartz cuvette. Wave-
lengths ranging from 190 to 250 nm were measured, with
a 0.2-nm step resolution and 100 nmÆmin
)1
speed. The
temperature was controlled by a PTC-343 controller.
Spectra were collected and averaged over four scans. The
a-helical content was established from the amplitude at
222 nm, as previously described [32], assuming that only
a-helix and random coil conformations contribute to the
CD spectrum.
Structure calculation of penetratin
Distance constraints were generated from quantifying
NOESY (s
mix
¼ 100 ms) cross-peak intensities according

to the procedure previously outlined [25]. The structures
were checked against the NOESY data to verify that no
short distances had missing NOE cross-peaks in the data.
However, this procedure was of limited use due to overlap
with strong lipid cross-peaks. A total of 129 distance
constraints were identified, mostly sequential and medium-
range NOEs defining the secondary structure of the peptide
3056 M. Lindberg et al. (Eur. J. Biochem. 270) Ó FEBS 2003
(40 intraresidue, 54 sequential and 35 medium-range
NOEs). For a small peptide like penetratin one does not
expect to find many long-range NOEs. The structures were
calculated using
DYANA
[33] version 1.5, using the standard
annealing algorithm. A total of 60 structures were calculated
and 20 were selected to represent the final solution structure
based on their target function and constraint violations. The
quality of the structure was checked with the program
PROCHECK
_
NMR
[34] and analyses of secondary structure
were performed with the
GAP
software package [35]. The
structures were visualized using
INSIGHT
(version 2000,
Accelrys). The coordinates of the final structures together
with the input constraints have been deposited with the

PDB under accession number 1QMQ.
Results
Structure of penetratin in phospholipid bicelles
CD measurements were performed in order to establish the
effect of different phospholipid compositions on the secon-
dary structure of penetratin (Figs 1 and 2). Figure 1 shows
CD spectra from penetratin in different bicellar solutions.
The CD spectra reveal that the structure of penetratin is
mostly random coil in water (Fig. 1a) and that it interacts
strongly with the partly charged bicelles to obtain a helical
structure (Fig. 1d). There is also some structure induction
in neutral bicelles, although to much less extent (Fig. 1c).
Furthermore it is evident that the peptide does not interact
with DHPC alone to obtain a helical structure (Fig. 1b).
Therefore it is safe to assume that penetratin most likely
interacts with the charged surface of the negatively charged
bicelles while it does not interact with the DHPC rim. There
is a slight difference in the structure of penetratin in bicelles
containing 10% DMPG and 10% DMPS, and penetratin
seems to be more helical in bicelles with DMPG (43% vs.
35%, Fig. 2b). The effect of varying the lipid/peptide ratio
(L/P) was also investigated and it was seen that the amount
of helical structure increased with higher L/P ratio in
agreement with what has previously been observed in
phospholipid vesicles [36].
To investigate in more detail the structure of the peptide
in the bicelle solvent, a solution structure of penetratin was
calculated based on 129 distance constraints. Resonance
assignments for all but the two terminal residues were
obtained from analysis of TOCSY data and sequential

assignments were obtained from NOE connectivities
(Fig. 3). Structural statistics for the ensemble of 20 struc-
tures are presented in Table 1 and the structure represented
by an ensemble of 20 models is shown in Fig. 4. The
structure shows low constraint violations and good stereo-
chemical properties, with only one residue within the
entire ensemble falling in the disallowed region of the
Ramachandran plot.
Based on analyses of hydrogen bonds and backbone
torsion angles, helical structure could be assigned for
residues Lys46 through Met54, although Lys46 is not as
well-defined and shows weaker hydrogen bonds. Arg53, at
the C-terminus of the helix is involved in hydrogen bonds to
both Gln50 and Phe49, indicating a mixture of a-helical and
3
10
-helical character. Met54 is only hydrogen bonded to
Asn51. Although the helix is somewhat irregular, we are
nevertheless able to identify a helical segment in the central
part of penetratin. The amount of helix is greater in the
structure than that predicted from CD measurements, but
this can be explained by an equilibrium between penetratin
bound to bicelles and in water, where there is a fast
exchange between the two forms. The NMR structure is
representative for the bicelle-bound form of penetratin,
while CD spectra represent an average over the sample.
Structure of penetratin(W48F,W56F) in phospholipid
bicelles
The nonactive penetratin(W48F,W56F) variant was studied
by high resolution NMR in both phospholipid bicelles and

SDS micelles. Experiments and assignments were made as
described above for penetratin. Secondary chemical shifts
for the H
a
resonances in peptides or proteins carry
information on secondary structure [37,38] and they were
Fig. 1. CD-spectra for penetratin in different solvents. (a) in buffer; (b)
300 m
M
DHPC; (c) q ¼ 0.5 DMPC/DHPC bicelles; (d) q ¼ 0.5
DMPC/DMPG/DHPC bicelles ([DMPG]/[DMPC] ¼ 0.1). The tem-
perature was 37 °CandthepH5.5.
Fig. 2. Effect of surface charge and peptide concentration on the
a-helical content in CD-spectra of penetratin for different compositions
of the bicelles. In all samples, the ratio of long-chained to short-chained
phospholipids was q ¼ 0.5. The temperature was 37 °Candthe
pH 5.5. (a) [DMPG]/[DMPC] ¼ 0.1, 1 m
M
penetratin; (b) [DMPS]/
[DMPC] ¼ 0.1, 1 m
M
penetratin; (c) [DMPG]/[DMPC] ¼ 0.1, 3 m
M
penetratin; (d) [DMPG]/[DMPC] ¼ 0.05, 1 m
M
penetratin.
Ó FEBS 2003 Structure and position of penetratin (Eur. J. Biochem. 270) 3057
calculated for penetratin(W48F,W56F) as well as for
penetratin according to Sykes and coworkers in both
solvents (Fig. 5). The data on penetratin in SDS micelles

was taken from [10]. Analysis of the NOESY spectrum for
penetratin(W48F,W56F) provided characteristic amide-
amide, d
aN
(i,i +3) and d
ab
(i,i + 3) NOE cross-peaks
indicative of a-helical secondary structure. The NOE results
for penetratin(W48F,W56F) and penetratin in the phos-
pholipid bicelles are summarized in Fig. 6. Based on the
similarity of the chemical shift and NOE data, we conclude
that the structure of penetratin(W48F,W56F) should be
very similar to that of penetratin in bicelles.
The NOE cross-peak patterns for penetra-
tin(W48F,W56F) in SDS gave similar evidence of a-helix
Fig. 3. NMR data for penetratin in phospholipid bicelles (q = 0.5,
[DMPG]/[DMPC] = 0.1). (A)Partofa600-MHzTOCSYspectrum
recorded at 37 °CshowingtheH
N
–H
a
region; (B) the H
N
–H
N
region
of a 600-MHz NOESY spectrum (s
mix
¼ 100 ms) with the indicated
assignments.

Table 1. Structural statistics for the ensemble of 20 penetratin structures
in bicelles calculated with
DYANA
.
Number of constraints 129
Target function (A
˚
2
) 0.04
Maximum distance violation (A
˚
) 0.07
Backbone rmsd (A
˚
)
All residues 1.24
Residues 45–55 0.64
Ramachandran plot regions (%)
Most favored 76.4
Allowed region 18.2
Generously allowed 5.0
Disallowed 0.4
Fig. 4. Solution structure of penetratin in acidic phospholipid bicelles
representedbyanensembleof20structures.The overlay was performed
by superimposing backbone atoms in residues Ile45-Lys55.
Fig. 5. Secondary H
a
chemical shifts for penetratin and penetra-
tin(W48F,W56). (A) penetratin (filled squares) and penetra-
tin(W48F,W56) (open squares) in phospholipid bicelles with q ¼ 0.5

and [DMPG]/[DMPC] ¼ 0.1. (B) penetratin (filled circles) and
penetratin(W48F,W56F) (open circles) in 300 m
M
SDS. The data on
penetratin in SDS micelles was taken from [10].
3058 M. Lindberg et al. (Eur. J. Biochem. 270) Ó FEBS 2003
as seen for active penetratin (data not shown). The
secondary chemical shifts in SDS indicate that the peptide
adopts an a-helical conformation to a larger extent in this
solvent. Previous investigations of peptides in bicelles and
SDS have shown that SDS can restrain motional flexibility
of the peptide, which might lead to a more structured
peptide.
The position of penetratin with respect to the bicelle
Paramagnetic probes were added to the different samples
to determine positioning of the peptide relative to the
surface and interior of the bicelle and micelle, respectively.
In studies of SDS micelles alone, Mn
2+
ions have been
shown to affect SDS resonances from nuclei at the surface
of the micelle. 5-doxyl- and 12-doxyl-labeled stearic acids
were shown to affect resonances from nuclei inside
(5-doxyl) or deeply buried (12-doxyl) in the micelle [39].
Interpretation of the paramagnetic probe experiments can
be done semiquantitatively by evaluating the loss of
amplitude for the cross-peaks of the peptide, by measuring
the remaining amplitude. The remaining amplitude [40],
RA, is defined as:
RA ¼ N Á

A
paramag:
A
0
where A
paramag
is the amplitude of the crosspeak measured
when the paramagnetic agent is added and A
0
is the
amplitude with no paramagnetic agent present. N is a
normalizing factor in order to normalize the remaining
amplitude so that the least affected crosspeak has a
remaining amplitude of 100%. Similarly, the position of
the peptide relative to the bicelle can be estimated by
observing the effect of specific paramagnetic probes on
resonances in the NMR spectrum. In the present study we
have investigated the effects of Mn
2+
ions as well as of a
5-doxyl- and 12-doxyl-labeled phospholipid on the
1
H
resonances of the peptide. The 5-doxyl labeled phospholipid
has been shown to insert into phospholipid vesicles with the
doxyl group at a distance from the center of the bilayer of
12 A
˚
[41,42]. These measurements are only semiquantitative
and we judge that the errors on the remaining amplitudes

are on the order of ± 20%.
Looking at the lipids, the results show that the paramag-
netic probes clearly affect the resonances originating from
the phospholipids. The Mn
2+
ions very efficiently remove
resonances of protons close to the phosphate head-group,
i.e. of the choline CH
2
protons (4.30 p.p.m and 3.68 p.p.m),
and the 2-CH glycerol proton (at 5.26 p.p.m), while leaving
aliphatic side-chain protons unaffected (data not shown).
The 5-doxyl-group clearly affects signals originating from
the aliphatic side-chains, as well as glycerol lipid signals,
although no visible effect was seen on the methyl resonance.
The 12-doxyl-labeled phospholipid has a large broadening
effect on the methyl proton resonances as well as on the rest
of the aliphatic side-chain, while leaving the choline and
glycerol protons, close to the head-group, much less
broadened.
Curves depicting the effect of the paramagnetic probes
on signal intensities for penetratin in bicelles are shown
in Fig. 7. The addition of paramagnetic agents in low
concentrations does not seem to alter the structure of the
peptide or integrity of the bicelles as the penetratin
spectrum remained essentially the same. Overall, Mn
2+
ions do not seem to have a specific effect on the
penetratin resonances (Fig. 7A) and it is difficult to judge
whether there is a general line-broadening effect on the

entire peptide or if the peptide is more or less protected
from Mn
2+
ions. Resonances from side-chain protons
belonging to Gln44, Lys55 and Lys57 disappear in
the presence of Mn
2+
already at lower concentrations,
0.25–0.5 m
M
.
NOEs were observed between the H
N
resonances
belonging to residues Gln50, Asn51, Arg53 and Lys55
and the CH
2
choline protons (at 4.30 p.p.m., 3.68 p.p.m).
The side-chains of the two tryptophan residues, Trp48 and
Trp56 have NOEs to both the 2-CH
2
glycerol protons and
to the choline protons (3.68 p.p.m) indicating that the
peptide resides within the head-group region of the bilayer,
supporting the conclusion that the Mn
2+
ions have a slight
general broadening effect on the peptide.
The residues most affected by the 5-doxyl-labeled
phospholipid are Ile47, Trp48 and Phe49, situated at the

N-terminus, and surprisingly Met54. At a concentration of
1m
M
5-doxyl spin-label the cross-peaks for Trp48 and
Phe49 disappear while the cross-peak for Met54 is greatly
reduced (Fig. 7B), and when the spin-label concentration is
increased to 2 m
M
, the cross-peaks for residues Ile45
through Phe49 disappear completely.
The results obtained with the 12-doxyl labeled phospho-
lipid (Fig. 7C) were similar to those obtained with the lipid
labeled at position 5, i.e. the same residues were primarily
affected, although to a much less extent, and signal still
remained for Ile45 through Ile47 even at a concentration of
2m
M
.
Fig. 6. A summary of amide-amide, d
aN
(i,i +3), d
aN
(i,i +4) and
d
ab
(i,i + 3) NOE connectivities. (A) penetratin in q ¼ 0.5 bicelles with
[DMPG]/[DMPC] ¼ 0.1 (B) penetratin(W48F,W56) in q ¼ 0.5
bicelles with [DMPG]/[DMPC] ¼ 0.1.
Ó FEBS 2003 Structure and position of penetratin (Eur. J. Biochem. 270) 3059
Positioning of penetratin(W48F,W56F) in phospholipid

bicelles
The same experiments with paramagnetic probes was
performed for penetratin(W48F,W56F). The results are
summarized in Fig. 7 together with the results obtained for
penetratin, and it can be seen that the trends in the results
are similar. Notably, there is a significant effect of the Mn
2+
ions on the C-terminal residues Trp56, Lys57 and Lys58,
which is not seen in penetratin. In analogy with what was
discussed for penetratin, this implies that the N-terminal
residues are more protected from Mn
2+
ions than the
C-terminus, suggesting that there is a difference from the
broadening effect seen for penetratin.
The experiments using the 5-doxyl phospholipid
(Fig. 7B) show that cross peaks for residues Ile45–Glu50
and Met54 are broadened by the spin label, while most of
the C-terminal residues are much less affected. The results
from adding 12-doxyl to penetratin(W48F,W56F) (Fig. 7C)
show that cross-peaks from residues Ile47, Phe48 and Phe49
completely vanish. Ile45 and Lys46 are less affected but still
located near the paramagnetic doxyl group. The C-terminal
residues Met54–Lys58 are affected only marginally by this
spin label.
Positioning of penetratin(W48F,W56F) in SDS micelles
Experiments with penetratin(W48F,W56F) in SDS micelles
were performed, partly to compare the effects of the two
solvent systems and partly to compare with results obtained
previously for penetratin in SDS [10]. Figure 8 shows the

remaining amplitude of TOCSY H
N
-H
a
cross-peaks for
penetratin and penetratin(W48F,W56F) in SDS micelles
with Mn
2+
ions, 5-doxyl stearic acid and 12-doxyl stearic
acid added. These results are not as straightforward to
interpret as the results from bicelles. When Mn
2+
ions are
added to penetratin(W48F,W56F) in SDS, two residues are
more affected than the others, Phe56 and Lys58 (Fig. 8A).
At higher Mn
2+
concentrations (1.5 m
M
), residues in the
C-terminal part are more affected than what is seen in the
N-terminal part (data not shown).
The results for penetratin(W48F,W56F) with 5-doxyl
stearic acid show that the spin label has the largest effect on
Phe48 and the on three residues, Lys55, Phe56 and Lys57.
Thus it would seem that the same residues are more or less
affected by both Mn
2+
and 5-doxyl stearic acid. Finally,
adding 12-doxyl stearic acid to penetratin(W48F,W56F)

Fig. 7. The remaining amplitude of H
N
–H
a
cross-peaks in 600 MHz
TOCSY spectra recorded at 45 °Cfor3m
M
penetratin (closed) and
3m
M
penetratin(W48F,W56F) (open) in q = 0.5 bicelles with
[DMPG]/[DMPC] = 0.1. The paramagnetic agents are (A) 2 m
M
MnCl
2
(squares) (B) 1 m
M
1-palmitoyl-2-stearoyl-sn-glycero-5-doxyl-
3-phosphatidylcholine (circles), and (C) 1 m
M
1-palmitoyl-2-stearoyl-
sn-glycero-12-doxyl-3-phosphatidylcholine (diamonds).
Fig. 8. The remaining amplitude of H
N
–H
a
cross-peaks in 600 MHz
TOCSY recorded at 45 °Cfor3m
M
penetratin (closed) and 3 m

M
penetratin(W48F,W56F) (open) in SDS micelles. The paramagnetic
agents are (A) 0.5 m
M
MnCl
2
(squares) (B) 5 m
M
5-doxyl stearic acid
(circles), and (B) 5 m
M
12-doxyl stearic acid (diamonds).
3060 M. Lindberg et al. (Eur. J. Biochem. 270) Ó FEBS 2003
in SDS micelles shows that two residues, Phe49 and Met54
are most strongly affected.
Discussion
Penetratin adopts an a-helical structure between residues
Lys46 and Met54 in acidic bicellar solution; this structure is
similar to the secondary structure of penetratin in SDS
micelles [10,14]. Chemical shift and NOE data for pene-
tratin(W48F,W56F) indicate that the secondary structure is
very similar to the active penetratin in both SDS micelles
and phospholipid bicelles. These results suggest that the
secondary structure is conserved when the two tryptophan
residues are replaced with phenylalanines. The large differ-
ence in chemical shift observed for Phe49 in penetratin and
in penetratin(W48F,W56F) is most likely due to changes in
ring currents when changing amino acid 48 from a
tryptophan to a phenylalanine, an effect that is seen in
both bicelles and SDS micelles. Hence, we conclude that

replacing the two tryptophan residues by phenylalanines,
turning penetratin into a nontranslocating peptide does not
change the secondary structure.
Next, the paramagnetic broadening effects and hence the
positioning of the peptide relative to the surface and interior
of the phospholipid bicelle were studied for both peptides
(Fig. 7). For penetratin, it is difficult to judge from the line-
broadening caused by Mn
2+
ions whether the entire peptide
is protected from Mn
2+
, or that a more general broadening
effect is seen. However, the observed peptide–lipid NOEs,
together with the fact that Mn
2+
ions seem to affect the
head-group region of the lipids suggest that the peptide
resides within the phospholipid head-group layer, at the
interface between the head-group region and the hydro-
phobic interior. This is supported by the results obtained
with the two spin-labeled phospholipids, which show that
the hydrophobic residues at the N-terminus are positioned
towards the hydrophobic interior of the bicelle. This leads us
to believe that penetratin resides more or less parallel to the
bicelle surface with its hydrophobic residues interacting with
the interior of the bicelle. The results from the spin-label
study are mapped on the penetratin structure in Fig. 9,
where it is clear that the NOEs and spin-label results are
both consistent with the peptide being positioned within the

head-group region.
Interestingly, subtle differences can be observed for
penetratin(W48F,W56F) (Fig. 9). Mn
2+
ions have a more
specific effect on the C-terminal residues of this peptide than
on penetratin, which implies that part of the peptide is more
exposed to Mn
2+
. In addition the 12-doxyl labeled lipid
has a greater effect on the N-terminus than the 5-doxyl
lipid has, indicating that the N-terminus of pene-
tratin(W48F,W56F) inserts somewhat deeper into the
bilayer than what is seen for penetratin. This is consistent
with the Mn
2+
results and constitutes a significant differ-
ence between penetratin and penetratin(W48F,W56F).
The N-terminal residues in penetratin(W48F,W56F)
seem to be affected by all probes, but to a varying extent
indicating a great deal of flexibility, as also suggested by
the penetratin structure. In addition one must consider the
inherent flexibility of the phospholipids, which adds to the
uncertainty in the estimated position. In fact, the 5-doxyl
phospholipid and the Mn
2+
both affect protons in the head-
group region of the phospholipids, while the 12-doxyl does
not. Hence it is not surprising that there is overlap between
the areas that these probes measure. Nevertheless, we

conclude that both penetratin and penetratin(W48F,W56F)
interact with the bicelle surface, as shown by CD, but in
slightly different ways. The proposed mechanism for
penetratin translocation includes interactions between the
charged residues of penetratin and the membrane surface
layer, and by substituting the two tryptophans for phenyl-
alanines in penetratin, the peptide inserts somewhat deeper
into the hydrophobic bilayer, and looses the ability to
translocate.
Finally, comparison between the position of penetratin
and the analog in bicelles and in SDS reveals significant
differences. The SDS data are very complex and suggest that
the micelle is not a well-defined system in which the peptide
is positioned in a simple way. The micelle and its constit-
uents are highly flexible and it is clearly seen that several
nonterminal residues are affected by more than one
paramagnetic probe, especially at the C-terminus of the
peptides. These results show that one should be careful
when interpreting positioning data from SDS micelles
because the size and the curvature of the micelle might
force the peptide to a certain position. However, there are
also very different charge densities associated with the two
membrane mimetic solvents investigated here, which may be
important for the positioning experiments with the highly
charged peptides. These observations emphasize the import-
ance of using realistic membrane substitutes in studies of
membrane–peptide interactions, such as positioning inves-
tigations. Although the induced secondary structure seems
similar in the two solvents, the paramagnetic broadening
studies in SDS are much more difficult to interpret and do

not entirely agree with what was found in phospholipid
bicelles.
Conclusions
We have determined a solution structure of penetratin in
partly charged bicelles. A helical structure is seen for around
Fig. 9. The positioning results for penetratin and penetra-
tin(W48F,W56F) mapped onto the structure of penetratin. The lines
represent the average size of the head-group region of the bilayer. Only
effects that result in a remaining amplitude of <0.4 are shown. Blue
color indicates the effect of Mn
2+
ions, light red the effect of
1-palmitoyl-2-stearoyl-sn-glycero-5-doxyl-3-phosphatidylcholine, and
dark red the effect of 1 m
M
1-palmitoyl-2-stearoyl-sn-glycero-12-
doxyl-3-phosphatidylcholine.
Ó FEBS 2003 Structure and position of penetratin (Eur. J. Biochem. 270) 3061
50% of the peptide (Lys46–Met54). Our results show that
penetratin preferentially interacts with the bicellar surface
formed by the DMPC/DMPG lipids, as little structure
induction is seen in DHPC and in neutral bicelles. Penetr-
atin and penetratin(W48F,W56F) are structurally very
similar to each other when interacting with phospholipid
bicelles. Both peptides are positioned within the bicelle;
however, subtle differences in the positioning of the two
peptides are seen. Penetratin is not deeply inserted into the
lipid bilayer, but seems nevertheless to reside within the
bicelle head-group layer, with NOE data suggesting a
parallel orientation relative to the surface. The observations

further show that the N-terminus of penetratin
(W48F,W56F) inserts more deeply into the bicelle as
compared to penetratin. This in turn may be the result of
a different interaction between the peptide and membrane,
affecting its cell-penetrating properties. Furthermore, the
results clearly show that phospholipid bicellar solutions
provide suitable membrane substitutes for cell-penetrating
peptides, such as penetratin. An important advantage of
using phospholipid bicelles over detergent micelles is
illustrated by the fact that the position of the peptide
relative to the bicelle surface can be investigated in a realistic
membrane-like environment.
Acknowledgements
We thank the Swedish NMR Center, Gothenburg for access to the
NMR spectrometers. This study was supported by grants from the
Swedish Science Research Council (to L. M. and A. G.).
References
1. Derossi, D., Chassaing, G. & Prochiantz, A. (1998) Trojan pep-
tides: the penetratin system for intracellular delivery. Trends Cell
Biol. 8, 84–87.
2. Lindgren, M., Ha
¨
llbrink, M., Prochiantz, A. & Langel, U
¨
. (2000)
Cell-penetrating peptides. Trends Pharmacol. Sci. 21, 99–103.
3. Takeshima, K., Chikushi, A., Lee, K K., Yonehara, S. & Mat-
suzaki, K. (2003) Translocation of analogues of the antimicrobial
peptides magainin and buforin across human cell membranes.
J. Biol. Chem. 278, 1310–1315.

4. Qian,Y.,Billeter,M.,Otting,G.,Mu
¨
ller,M.,Gehring,W.J.&
Wu
¨
thrich, K. (1989) The structure of the Antennapedia homeo-
domain determined by NMR spectroscopy in solution: compari-
son with prokaryotic repressors. Cell 59, 573–580.
5. Derossi, D., Joliot, A., Chassing, G. & Prochiantz, A. (1994) The
third helix of the Antennapedia homeodomain translocates
through biological membranes. J. Biol. Chem. 269, 10444–10450.
6. Derossi, D., Calvet, S., Trembleau, A., Brunissen, A., Chassaing,
G. & Prochiantz, A. (1996) Cell internalization of the third helix of
the Antennapedia homeodomain is receptor-independent. J. Biol.
Chem. 271, 18188–18193.
7. Prochiantz, A. (1999) Homeodomain-derived peptides: in and out
of the cells. Ann. N. Y. Acad. Sci. 886, 172–179.
8. Thore
´
n, P., Persson, D., Karlsson, M. & Norde
´
n, B. (2000) The
Antennapedia peptide penetratin translocates across lipid bilayers
– the first direct observation. FEBS Lett. 482, 265–268.
9. Drin, G., De
´
me
´
ne
´

, H., Temsamani, J. & Brasseur, R. (2001)
Translocation of the pAntp Peptide and Its Amphipathic Ana-
logue AP-2AL. Biochemistry 40, 1824–1834.
10. Lindberg, M. & Gra
¨
slund, A. (2001) The position of the cell
penetrating peptide penetratin in SDS micelles determined by
NMR. FEBS Lett. 497, 39–44.
11. Magzoub, M., Kilk, K., Eriksson, L.E., Langel, U. & Gra
¨
slund,
A. (2001) Interaction and structure induction of cell-penetrating
peptides in the presence of phospholipid vesicles. Biochim. Bio-
phys. Acta 1512, 77–89.
12. Persson, D., Thore
´
n, P.E.G. & Norde
´
n, B. (2001) Penetratin-
induced aggregation and subsequent dissociation of negatively
charged phospholipid vesicles. FEBS Lett. 505, 307–312.
13. Prochiantz, A. (1996) Getting hydrophilic compounds into cells:
lessons from homeopeptides. Curr. Opin. Neurobiol. 6, 629–634.
14. Berlose,J.P.,Convert,O.,Derossi,D.,Brunissen,A.&Chassaing,
G. (1996) Conformational and associative behaviours of the third
helix of antennapedia homeodomain in membrane-mimetic
environments. Eur. J. Biochem. 242, 372–386.
15. Scheller,A.,Wiesner,B.,Melzig,M.,Bienert,M.&Oehlke,J.
(2000) Evidence for an amphipathicity independent cellular
uptake of amphipathic cell-penetrating peptides. Eur. J. Biochem.

267, 6043–6049.
16. Vold, R. & Prosser, S. (1996) Magnetically oriented phospholipid
bilayered micelles for structural studies of polypeptides. Does the
ideal bicelle exist? J. Magn. Reson. 113, 267–271.
17. Struppe, J. & Vold, R.R. (1998) Dilute bicellar solutions for
structural NMR work. J. Magn. Reson. 135, 541–546.
18. Sanders, C.R. & Schwonek, J.P. (1992) Characterization of
magnetically orientable bilayers in mixtures of dihexanoylphos-
phatidylcholine and dimyristoylphosphatidylcholine by solid-state
NMR. Biochemistry 31, 8898–8905.
19. Struppe, J., Whiles, J.A. & Vold, R.R. (2000) Acidic phospholipid
bicelles: a versatile model membrane system. Biophys. J. 78,
281–289.
20. Luchette, P.A., Vetman, T.N., Prosser, R.S., Hancock, R.E.W.,
Nieh, M.P., Glinka, C.J., Krueger, S. & Katsaras, J. (2001)
Morphology of fast-tumbling bicelles: a small angle neutron
scattering and NMR study. Biochim. Biophys. Acta 1513, 83–94.
21. Glover, K.J., Whiles, J.A. & Wu, G., YuN J., Deems, R., Struppe,
J.O., Stark, R.E., Komives, E.A. & Vold, R.R. (2001) Structural
evaluation of phospholipid bicelles for solution-state studies of
membrane-associated biomolecules. Biophys. J. 81, 2163–2171.
22. Vold, R.R., Prosser, S.R. & Deese, A.J. (1997) Isotropic solutions
of phospholipid bicelles: a new membrane mimetic for high-
resolution NMR studies of polypeptides. J. Biomol. NMR 9,
329–335.
23. Whiles, J.A., Brasseur, R., Glover, K.J., Melacini, G., Komives,
E.A. & Vold, R.R. (2001) Orientation and effects of mastoparan X
on phospholipid bicelles. Biophys. J. 80, 280–293.
24. Chou, J.J., Kaufman, J.D., Stahl, S.J., Wingfield, P.T. & Bax, A.
(2002) Micelle-induced curvature in a water-insoluble HIV-1 Env

peptide revealed by NMR dipolar coupling measurement in
stretched polyacrylamide gel. J. Am. Chem. Soc. 124, 2450–2451.
25. Andersson, A. & Ma
¨
ler, L. (2002) NMR solution structure and
dynamics of motilin in isotropic phospholipid bicellar solution.
J. Biomol. NMR 24, 103–112.
26. Glover, K.J., Whiles, J.A., Vold, R.R. & Melacini, G. (2002)
Position of residues in transmembrane peptides with respect to the
lipid bilayer: a combined lipid NOEs and water chemical exchange
approach in phospholipid bicelles. J. Biomol. NMR 22, 57–64.
27. Israelachvili, J. (1991) Intermolecular Surface Forces,p.372.
Academic Press, San Diego.
28. States, D.J., Haberkorn, R.A. & Ruben, D.J. (1982) A two-
dimensional nuclear Overhauser experiment with pure absorption
phase in four quadrants. J. Magn. Reson. 48, 286–292.
29.Jeener,J.,Meier,B.,Bachman,P.&Ernst,R.R.(1979)
Investigation of exchange processes by two-dimensional NMR
spectroscopy. J. Chem. Phys. 71, 4546–4563.
30. Braunschweiler, L. & Ernst, R.R. (1983) Coherence transfer by
isotropic mixing: application to proton correlation spectroscopy.
J. Magn. Reson. 53, 521–528.
3062 M. Lindberg et al. (Eur. J. Biochem. 270) Ó FEBS 2003
31. Piotto, M., Saudek, V. & Sklena
´
r, V. (1992) Gradient-tailored
excitation for single-quantum NMR spectroscopy of aqueous
solutions. J. Biomol. NMR 2, 661–665.
32. Greenfield, N. & Fasman, G.D. (1969) Computed circular
dichroism spectra for the evaluation of protein conformation.

Biochemistry 8, 4108–4116.
33. Gu
¨
ntert, P., Mumenthaler, C. & Wu
¨
thrich, K. (1997) Torsion
angle dynamics for NMR structure calculations with the new
program DYANA. J. Mol. Biol. 273, 283–298.
34. Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton,
J.M. (1993) PROCHECK: a program to check the stereochemical
quality of protein structures. J. Appl. Crystallogr. 26, 283–291.
35. Gippert, G. (1995) New computational methods for 3D NMR data
analysis and protein structure determination in high-resolution
internal coordinate space. PhD Thesis, The Scripps Research
Institute, La Jolla, CA.
36. Magzoub, M., Eriksson, L.E.G. & Gra
¨
slund, A. (2002) Con-
formational states of the cell-penetrating peptide penetratin
when interacting with phospholipid vesicles: effects of surface
charge and peptide concentration. Biochim. Biophys. Acta 1563,
53–63.
37. Wishart, D.S., Sykes, B.D. & Richards, F.M. (1991) Relationship
between nuclear magnetic resonance chemical shift and protein
secondary structure. J. Mol. Biol. 222, 311–333.
38. Wishart, D.S. & Sykes, B.D. (1994) Chemical shifts as a tool for
structure determination. Methods Enzymol. 239, 363–392.
39. Damberg, P., Jarvet, J. & Gra
¨
slund, A. (2001) Micellar systems as

solvents in peptide and protein structure determination. Methods
Enzymol. 339, 271–285.
40. Lindberg, M., Jarvet, J., Langel, U
¨
.&Gra
¨
slund, A. (2001)
Secondary structure and position of the cell-penetrating peptide
transportan in SDS micelles as determined by NMR. Biochemistry
40, 3141–3149.
41. Chattopadhyay, A. & London, E. (1987) Parallax method for
direct measurement of membrane penetration depth utilizing
fluorescence quenching by spin-labeled phospholipids. Biochem-
istry 26, 39–45.
42. Abrams, F.S. & London, E. (1993) Extension of the parallax
analysis of membrane penetration depth to the polar region of
model membranes: Use of fluorescence quenching by a spin-label
attached to the phospholipid polar headgroup. Biochemistry 32,
10826–10831.
Ó FEBS 2003 Structure and position of penetratin (Eur. J. Biochem. 270) 3063