Comparing lipid photo-cross-linking efficacy of penetratin analogues bearing three different photoprobes: Dithienyl ketone, benzophenone, and trifluoromethylaryldiazirine

Article abstract of DOI:10.1021/bc900466q

Photoactivatable penetratin analogues bearing three different photoprobes, which do not disturb the membranotropic properties of the peptides, have been tested for their photo-cross-linking efficacy in glycerol and lipid media. In the case of glycerol, photo-cross-linking was observed, whereas in the case of SDS (used as a membrane model system), the dynamics of the SDS/penetratin assemblies and the photosensitizer properties of the probes prevented the cross-linking between the peptide and SDS. Bilayers of DMPG were partially photo-cross-linked by the penetratin analogues containing either a benzophenone or a trifluoromethylaryl-diazirine, whereas dithienyl ketone acted exclusively as a photosensitizer. The characterization by MALDI-TOF mass spectrometry of the photoadducts formed after irradiation required basic hydrolysis of DMPG for an efficient capture of the biotinylated peptide analogues with streptavidin-coated magnetic beads. MALDI-TOF analysis of the photoadducts between the photoactivatable penetratin and DMPG allowed an unambiguous identification of the covalent bond formed with the lipids. Altogether, we show herein that the efficacy of the lipid photo-cross-linking depends on the environment, the dynamics of the supramolecular assembly, and the physicochemical properties of the photoprobe.

Full text of DOI:10.1021/bc900466q

352
Bioconjugate Chem. 2010, 21, 352–359
Comparing Lipid Photo-Cross-linking Efficacy of Penetratin Analogues
Bearing Three Different Photoprobes: Dithienyl Ketone, Benzophenone, and
Trifluoromethylaryldiazirine
Chen-Yu Jiao, Isabel D. Alves, Vanessa Point, Solange Lavielle, Sandrine Sagan, and Ge´rard Chassaing*
UPMC Paris 06 - ENS - CNRS, UMR 7203, Laboratoire des BioMole´cules, FR 2769 Chimie Mole´culaire, Universite´ Pierre et
Marie Curie, Case courier 182, 4, Place Jussieu, 75252 Paris Cedex 05, France. Received October 26, 2009;
Revised Manuscript Received January 4, 2010
Photoactivatable penetratin analogues bearing three different photoprobes, which do not disturb the membranotropic
properties of the peptides, have been tested for their photo-cross-linking efficacy in glycerol and lipid media. In
the case of glycerol, photo-cross-linking was observed, whereas in the case of SDS (used as a membrane model
system), the dynamics of the SDS/penetratin assemblies and the photosensitizer properties of the probes prevented
the cross-linking between the peptide and SDS. Bilayers of DMPG were partially photo-cross-linked by the
penetratin analogues containing either a benzophenone or a trifluoromethylaryl-diazirine, whereas dithienyl ketone
acted exclusively as a photosensitizer. The characterization by MALDI-TOF mass spectrometry of the photoadducts
formed after irradiation required basic hydrolysis of DMPG for an efficient capture of the biotinylated peptide
analogues with streptavidin-coated magnetic beads. MALDI-TOF analysis of the photoadducts between the
photoactivatable penetratin and DMPG allowed an unambiguous identification of the covalent bond formed with
the lipids. Altogether, we show herein that the efficacy of the lipid photo-cross-linking depends on the environment,
the dynamics of the supramolecular assembly, and the physicochemical properties of the photoprobe.
plexes such as benzophenone/cyclodextrin or transmembrane
peptide/phospholipids in which the affinity is low (t1/2 value
around milliseconds), the irradiation-assisted cross-linking was
not inhibited (7, 8). The confinement of the reactive species
and the “freezing” of internal movements of both the ligand
and the receptor (or guest molecule) within these complexes
restrain the intrinsic mobility of both molecules. Thus, it is
believed that the covalent bond formation likely represents a
snapshot of the interaction.
INTRODUCTION
The formation of a covalent bond, after irradiation of a
photoactivatable molecule, is a commonly used strategy for
linking ligands to both hard and soft surfaces (1-3). For that
purpose, four types of photoactivatable aryl reagents are mainly
used: aryldiazonium salts, diazoaryl, trifluoromethylaryldiazir-
ines, and phenylketone (4). To study by photolabeling the
interaction topography between a ligand and its protein partner,
the specifications are more stringent, since the photoactivatable
probe should be located in the immediate vicinity of the binding
domain and, upon irradiation, should form promptly a covalent
bond with the protein (5). The bioconjugated construct possesses
different “moieties”. A peptide ligand usually includes an
adequately positioned photoprobe in its sequence, together with
a recognition moiety (e.g., biotin) used to purify the covalent
complex from the incubation mixture (6). The tricky step is
often the correct rational design of a ligand that retains a high
and specific affinity for the protein partner.
Upon irradiation, the photoactivatable probe is transformed
into a highly reactive radical species that can abstract a hydrogen
atom present in its close environment within the protein binding
site. Thus, an important aspect to consider is the reactivity of
the chosen photoreactive probe, which must lead to molecular
species reactive enough within the time scale of the ligand-
protein complex half-life time. Generally, the binding affinity
of peptide ligands for their receptor is high (in the nanomolar
range), corresponding to formation of complexes with a half-
life time in the minutes range. This corresponds to a much longer
time than what is mostly observed for reactive species (t1/2 values
around microseconds or shorter). Even in “host-guest” com-
The goal of the present study was to test the feasibility and
the limitations of photochemistry in the analysis of the interac-
tion(s) of a cationic peptide with a highly dynamic system, such
as plasma membranes constituted of a large range of molecular
components and supramolecular assemblies. Plasma membranes
are considered a mosaic of coexisting domains: the fluid lipid
matrix in disordered phase, lipid ordered domains (rafts),
nanodomains associated with different types of proteins (ca-
veolins, clathrin, actin, receptors, protein channels, etc.) and
some transient confinement zones (9, 10). The overall dynamic
of this mosaic is considered to occur in the nano- to milliseconds
time scale (11). However, if one considers the rotation around
a C-C bond or along the membrane axis and the aliphatic chain
wobbling in the disordered phase, the correlation times of
phospholipids greatly vary from 10-12 to 10-7 s. The correlation
time for the lateral diffusion is greater than 10-5 s and
supramolecular assemblages occur starting from very small rafts
(diameter <10 nm) to larger ones (from 30 to 230 nm) in time
scales of 10-3 s (12, 13). Photoactivatable lipids have been used
to probe the interaction within lipids (14). Modifications of lipids
by highly reactive radicals, such as oxygen-derived species, have
also been reported (15). Herein, we address the question of the
photochemistry strategy as a reliable method to study peptide/
lipid interaction.
* Corresponding author. Dr. Ge´rard Chassaing, UMR 7203, Labo-
ratoire des BioMole´cules, Case courier 182, 4, Place Jussieu, 75252
Paris Cedex 05, France. Fax: +33 1 44 27 71 50. Tel: +33 1 44 27 32
61. E-mail: gerard.chassaing@upmc.fr.
Penetratin, (corresponding to helix-3 of Antennapedia ho-
meodomain: Arg-Gln-Ile-Lys-Ile-Lys-Phe-Gln-Asn-Arg-Arg-
Met-Lys-Trp-Lys-Lys), is a cationic membranotropic peptide,
10.1021/bc900466q  2010 American Chemical Society
Published on Web 01/20/2010

Lipid Photo-Cross-Linking with Penetratin Analogues
Bioconjugate Chem., Vol. 21, No. 2, 2010 353
Scheme 1. Structures of the Three Photoactivatable Probes (X ) Bz, DTK, or TMD) Attached to Lys48 within the Sequence of
Penetratina
a The three penetratin analogues are modified at the N-terminus by Biotinylated-(Gly)4 and are abbreviated herein as Pen(Lys48Bz), Pen(Lys48DTK),
and Pen(Lys48TMD).
a prototype of cell-penetrating peptide, which is able to enter
eukaryotic cells without cytotoxity, and which presents anti-
microbial activities in some prokaryotic cells (16, 17). Even if
parallel orientations of penetratin to bilayer were clearly
established by various biophysical methods, a direct transloca-
tion has been shown as one of the entry mechanism of penetratin
in eukaryotic cells (18). As Trp48 was found located in the
hydrophobic area of phospholipids (19, 20), we have substituted
this residue by an acylated lysine on the side-chain with three
different photoprobes: phenylketone, dithienylketone, or trif-
luoromethylaryldiazirine (Scheme 1). As we have already
described for other studies (18, 21-23), these photoactivatable
penetratin analogues were also functionalized with biotin,
separated from the penetratin sequence by a spacer of four
glycines, which allows an easy purification through streptavidin-
grafted magnetic beads before MALDI-TOF mass spectrometry
analysis. The influence of the different photoprobes on the
membranotropic properties of penetratin has been analyzed by
monitoring their effect on lipid phase transition using differential
scanning calorimetric (DSC). Since electrostatic interactions
have been shown to play an important role in lipid interaction
with penetratin (24, 25) large multilamellar vesicles (MLVs)
made of dimyristoylphosphatidylglycerol (DMPG) were used
as models of the highly dynamic cell membrane. The photo-
chemical behavior of the different peptide analogues has been
analyzed in water, glycerol, and sodium dodecyl sulfate (SDS)
as models of aqueous, polar headgroups and lipophilic environ-
ment of the membrane, respectively. DMPG MLVs were
irradiated with the photoactivatable analogues (above the phase
transition temperature) and the resulting photoadducts analyzed
by mass spectrometry to compare their cross-linking efficacy.
alcohol and 4-(trifluoroacetyl)benzoic acid using anhydrous
magnesium sulfate and sulfuric acid. To a suspension of
MgSO4 (92 mmol) in CH2Cl2 (100 mL), H2SO4 (23 mmol)
was added dropwise After 15 min of stirring, tert-butyl
alcohol and 4-(trifluoroacetyl)benzoic acid were added. The
mixture was stirred for two days and the reaction was
quenched by addition of saturated NaHCO3 (170 mL). The
extraction with CH2Cl2 led to 10 mmol of pure 4-(trifluoro-
acetyl)benzoic tert-butyl ester. RF (cyclohexane/ethyl actetate,
1
8/2) ) 0.3. H NMR (CDCl3), δ: 1.62 (s, 9H), 8.13 (s, 4H).
The unreacted starting material was recovered after acidifica-
tion of the aqueous phase and extraction with CH2Cl2.
4-(2-Trifluoroethanoneoxime)benzoic tert-Butyl Ester 2. To
a stirred mixture of hydroxylamine hydrochloride (8 mmol) and
NaOH (8 mmol) in 33 mL of boiling ethanol, compound 1 (8
mmol in EtOH, 4 mL) was added dropwise. After refluxing for
3 days, ether was added, and the organic layers were washed
with 0.01 N HCl (3×) and with water (3×). RF (CHCl3) ) 0.25.
1H NMR (CDCl3), δ: 1.6 (s, 9H), 7.59 (d, 2H), 8.10 (q, 2H),
8.38 (s, broad 1H).
4-(2-Trifluoroethanone-O-(p-tolylsulfonyl)oxime)benzoic tert-
Butyl Ester 3. A mixture of compound 2 (19 mmol) and of
p-sulfonyl chloride (38 mmol) in dry pyridine (30 mL) was
refluxed for 3 days. After solvent evaporation, the pure product
(yield 73%) was obtained after flash chromatography, (cyclo-
hexane/ethyl acetate, 95/5). RF (CHCl3) ) 0.66. 1H NMR
(CDCl3), δ: 1.6 (s, 9H), 2.48 (s, 3H), 7.41 (t, 4H), 7.86 (d, 2H),
8.06 (d, 2H).
4-(Trifluoromethyldiaziridine)benzoic tert-Butyl Ester 4. Com-
pound 3 (2 mmol) was introduced in a Parr Instrument of 0.3
L capacity, then liquid ammonia (20 g) was condensed after
cooling the autoclave in liquid nitrogen for 15 min. The mixture
was stirred for two days. After evaporation, the pure product
(yield 88%) was obtained after flash chromatography using
EXPERIMENTAL PROCEDURES
DMPG was purchased from Genzyme (Switzerland) and was
used without further purification. 4-Carboxybenzophenone and
4-(trifluoroacetyl)benzoic acid were purchased from Aldrich and
Apollo Scientific Limited, respectively. The 5-carboxy-5′-
methyl-2,2′-dithienylketone was a generous gift, from the
“Laboratoire d’Inge´nierie Mole´culaire et Biochimie Pharma-
cologique”, Universite´ Paul Verlaine, Metz, France.
Syntheses of Photoactivatable Probes. 4-[1-(trifluorometh-
yl)-3H-diazirinin-3-yl]benzoic Acid. The synthesis of 4-[1-
(trifluoromethyl)-3H-diazirinin-3-yl]benzoic acid was based on
a series of reactions already used for this type of photoprobes
(26, 27).
1
CH2Cl2. RF (CHCl3) ) 0.5. H NMR (CDCl3), δ: 1.6 (s, 9H),
2.25 (d, 1H), 2.80 (d, 1H), 7.58 (d, 2H), 8.05 (d, 2H).
4-[1-(Trifluoromethyl)-3H-diazirinin-3-yl]benzoic tert-Butyl
Ester 5. Compound 4 (1 mmol) was mixed with freshly prepared
Ag2O (5 mmol). After stirring for 3 h in the dark, the mixture
was filtered on Celite and purified by flash chromatography
1
using cyclohexane. RF (CHCl3) ) 0.75. H NMR (CDCl3), δ:
1.59 (s, 9H), 7.24 (d, 2H), 8.01 (d, 2H).
4-[1-(Trifluoromethyl)-3H-diazirinin-3-yl]benzoic Acid 6. Com-
pound 5 (10 mmol) was stirred with TFA (50 mmol) in CH2Cl2
(10 mL). The reaction was completed after two hours stirring,
leading after usual workup to compound 6. RF (CHCl3) ) 0.2.
1H NMR (DMSO), δ: 7.40 (d, 2H), 8.06 (d, 2H), 13.34 (s, 1H).
4-(trifluoroacetyl)benzoic tert-Butyl Ester 1. 4-(Trifluoro-
acetyl) benzoic tert-butyl ester 1 was prepared from tert-butyl

354 Bioconjugate Chem., Vol. 21, No. 2, 2010
Jiao et al.
Synthesis of Penetratin Analogues. The penetratin analogues
[Lys48(Bz)]Penetratin, [Lys48(DTK)]Penetratin, and [Lys48(TMD)]-
Penetratin abbreviated as Pen(Lys48Bz), Pen(Lys48DTK), and
Pen(Lys48TMD), respectively, and modified on Lys48 (number
48 corresponding to the Trp position in the sequence of the
Antennapedia homeodomain (28) were synthesized by solid-
phase methodology on a ABI 431 Synthesizer, using either a
Rink amide resin or Rink amide MBHA resin by Fmoc or Boc
strategies, respectively. Lys48 side chain was protected either
by Fmoc or Dde protecting groups for the Boc and Fmoc
strategies, respectively. After selective removal of the protect-
ing group, the free amine of Lys48 was acylated either by
4-carboxybenzophenone, 5-carboxy-5′-methyl-2,2′-dithienylke-
tone or 4-[3-(trifluoromethyl)-3H-diazirinin-3-yl]benzoic acid,
using DCC/HOBt for the activation. The subsequent Ile45 was
then introduced manually. The other amino acids were coupled
on the synthesizer using standard protocols of Boc and Fmoc
chemistries and amino acid activation with DCC/HOBt in NMP.
After removal of the last protecting group from the N-terminal
glycine, the peptides were biotinylated in NMP, as already
described (18, 21-23). The resins were washed with NMP,
CH2Cl2, and MeOH and dried under vacuum. The peptides were
cleaved from the resin with TFA/H2O/TIS (Fmoc strategy) (95/
2.5/2.5) or HF/anisole (Boc strategy). The crude peptides were
lyophilized and purified by RP-HPLC. The peptides were
characterized by MALDI-TOF mass spectrometry leading to
m/z 2864 for Pen(Lys48Bz), 2908 for Pen(Lys48DTK), and
2886 for Pen(Lys48TMD).
used for the experiments performed in SDS, excepted that no
Triton was added during the incubation step.
MALDI-TOF Mass Spectrometry Analysis. The photoad-
ducts were eluted from the magnetic beads with the 2,5-
dihydroxybenzoic acid (DHB) matrix, in 1:1 (v/v) CH3CN/H2O
(1% phosphoric acid) used for MALDI analysis. After 10 min
incubation, 1 µL of bead-free supernatant was deposited on the
sample holder.
Quantification by MALDI-TOF Mass Spectrometry. After
irradiation in glycerol of deuterated Pen(Lys48Bz) (penetratin
analogue with four deuterated glycines), a known amount (2
nmol) of hydrogenated Pen(Lys48Bz) (with four 1H-glycines)
was added to the medium. The photoadduct was then purified
by incubating the mixture with streptavidin-coated magnetic
beads for 30 to 60 min under gentle agitation. After elution of
all biotinylated peptides (photoadduct or free peptide) with the
matrix as described above, 1 µL of bead-free supernatant was
1
deposited on the sample holder. The ratio of the H/2H MH+
m/z signals allowed the quantitative determination of the starting
peptide that remained intact after irradiation.
RESULTS
Hydrophobic Character of Penetratin Analogues. The
modification of the hydrophobic/hydrophilic balance induced
by the photoprobes on the behavior of these penetratin analogues
can be related to the concentration of acetonitrile required to
elute the corresponding peptides by HPLC (in mixture of
acetonitrile water 0.1% of TFA). The concentrations required
correspond to 32.4% CH3CN for Pen(Lys48Bz), 21.5% CH3CN
for Pen(Lys48DTK), and 32% CH3CN for Pen(Lys48TMD),
leading to the following hydrophobic scale: Pen(Lys48DTK) >
Pen(Lys48Bz) ≈ (Lys48TMD).
Differential Scanning Calorimetry. The calorimetry was
performed on a high-sensitivity differential scanning calorimeter
(Calorimetry Sciences Corporation). A scan rate of 1 °C/min
was used, and there was a delay of 10 min between sequential
scans in a series that allow for thermal equilibration. Data
analysis was performed with the fitting program Cpcalc provided
by CSC and plotted with Igor. DMPG (1 mg/mL) was used at
peptide/lipid ratio of 1/25. Samples containing the peptide alone,
dissolved in buffer at similar peptide concentrations, exhibited
no thermal events over the temperature range 0-100 °C. This
indicates that the peptides do not denature over this temperature
range and that the endothermic events observed in this study
arise solely from phase transitions of the phospholipid vesicles.
A minimum of at least three to four heating and cooling scans
were performed for each analysis depending on whether or not
the spectra was reproducible.
Effect of the Medium on the Photochemical Properties
of the Different Photoprobes. Ammonium acetate buffer,
glycerol, and sodium dodecyl sulfate (SDS) were used as models
of the membrane interface. The crude products obtained after
irradiation (1 h) at 365 nm of these photoactivatable penetratin
analogues within one of the media were analyzed by MALDI-
TOF mass spectrometry. As already observed for peptides
containing benzophenone, the Pen(Lys48Bz) analogue was
mainly unaffected by irradiation in ammonium acetate buffer.
Pen(Lys48DTK) was strongly oxidized as a consequence of the
irradiation, MALDI-TOF spectra showing the incorporation of
one to four oxygen atoms. Pen(Lys48TMD) was nearly totally
oxidized but with the incorporation of a single oxygen atom
(+16).
Photolabeling. The sample of Pen(Lys48Bz), Pen(Lys48DTK),
or Pen(Lys48TMD) dissolved at micromolar concentration in
water or glycerol (500 µL) was irradiated on ice for 1 h at 365
nm with an UV light (HPR 125-W) located between 6 to 10
cm from the Eppendorf tube corresponding to a surface area of
0.39 cm2. Then, 1 µL of the irradiated sample was deposited
on the sample holder for direct MALDI-TOF mass spectrometry
analysis.
DMPG (1 µM in 20 mM ammonium acetate buffer at pH
7.5) was mixed with Pen(Lys48Bz), Pen(Lys48DTK), or
Pen(Lys48TMD) at 1/10 peptide/lipid ratio. Theses samples
were irradiated on ice for 1 h, under the conditions described
above, followed by a basic hydrolysis in MeOH/H2O (50/50)
using NaOH (0.1 M) at 37 °C overnight. After neutralization
with acetic acid, the methanol was evaporated before affinity
purification.
The polar heads of phospholipids represent a borderline
between two media, with a high dielectric constant (ε ) 80,
i.e., the aqueous phase) and a low dielectric constant (ε ) 3,
i.e., the lipid phase). This interface can be mimicked by glycerol
(ε ) 40). Glycerol adducts were efficiently formed with
Pen(Lys48Bz), since a signal with a mass increase of 92 was
observed. The signal intensity of the photoadduct was shown
to increase as a function of the irradiation time (Figure 2). The
disappearance as a function of time of the starting peptide
Pen(Lys48Bz) (bearing four deuterated glycines) was quantified
by addition after the irradiation of an internal standard (2 pmol)
of Pen(Lys48Bz) (bearing four hydrogenated glycines) (21). The
intact peptide remaining after irradiation was measured by
1
2
integrating the H and H m/z signals (2864.3 and 2872.4,
respectively), revealing an almost linear decrease of the starting
material (Figure 2) as a function of the irradiation time. In the
case of Pen(Lys48TMD), a mass increase of only 90, instead
of 92, was observed, which probably resulted from the alcohol
oxidation of the glycerol adduct. For Pen(Lys48TMD), the
starting peptide and the photoadduct coexisted with the mono-
anddioxidizedforms.WeaksignalsweredetectedforPen(Lys48DTK)
Purification with Streptavidin-Coated Magnetic Beads.
The hydrolyzed sample containing 0.1% of Triton was incubated
with streptavidin-coated magnetic beads (100 µg, Dynabeads
M280, Dynal) for 30 to 60 min under gentle agitation. The
washing steps (5× each) were performed with (1) buffer (Tris
pH 7, 50 mM) containing 0.1% SDS, (2) 50 mM Tris pH 7, (3)
NaCl 1 M, and (4) H2O/ACN 50/50. The same procedure was

Lipid Photo-Cross-Linking with Penetratin Analogues
Bioconjugate Chem., Vol. 21, No. 2, 2010 355
transition occurring at phase transition temperatures Tp and Tm
(29). The peak area corresponds to the enthalpy of the phase
transitions and the ∆H/Tm ratio to the entropy change. The main
phase transition is related to the chain melting during the gel-
to-liquid-crystalline phase change, whereas the pretransition
reflects changes in headgroup tilting and organization.
In agreement with previous studies, the addition of the
cationic peptide, penetratin, to anionic DMPG phospholipids
suppressed the pretransition temperature and decreased both the
enthalpy and the phase transition temperature (Tm) of the main
lipid transition (30). One would expect that the introduction of
hydrophobic probes in penetratin might promote the insertion
of the peptide into the lipid fatty acid chains and so lead to an
increased perturbation in the lipid phase transition. This was
not the case; indeed, the photoactivatable probes did not
significantly change the Tm and overall led to a smaller
perturbation in the main phase transition compared to the
original penetratin peptide. When comparing the three
photoactivatable analogues, Pen(Lys48DTK) perturbed least
the pretransition (that was not completly abolished) and main
phase transition, leading to a very small reduction in the
enthalpy of transition (Table 1). Altogether, the results dem-
onstrate that the introduction of the photoactivatable probes in
penetratin did not perturb the overall membranotropic properties
of the peptide.
Figure 1. High-sensitivity DSC heating scans. The spectra illustrate
the effect of adding penetratin (- · -) and penetratin photoactivatable
analogues: Pen(LysBz) ( · · · · ), Pen(LysDTK) (- -), and Pen(LysT-
MD) (s) on the thermotropic phase behavior of DMPG multilamellar
vesicles (s bold) at P/L of 1/25. Thermodynamic parameters are
given in Table 1.
with a limited number of oxidized states compared to those
observed in aqueous solution.
Finally, the reactivity of the photoactivatable penetratin
analogues in a lipophilic medium was mimicked by photoirra-
diation of the peptides in SDS solution. In that environment,
very low levels of cross-linking were detected, regardless of
the photoprobe used. A small photoadduct signal was only
observed for Pen(Lys48TMD). However, a high percentage of
peptide oxidation was observed after irradiation under these
conditions.
Analysis of the Photoadducts Obtained after Irradiation
of Peptides with Phospholipid Bilayers. No photoadduct was
observed by mass spectrometry after irradiation of DMPG
vesicles in the presence of photoactivatable penetratin analogues,
both using a crude preparation and after the purification step
with streptavidin-coated magnetic beads. In order to observe
photoadducts, the lipid moiety (hydrophobic character) of the
adducts had to be removed by basic hydrolysis, which was
performed overnight with NaOH (0.1 mol/L) in MeOH/H2O (50/
50) at 37 °C, followed by neutralization with acetic acid.
MALDI-TOF mass spectrometry analysis was performed after
a subsequent capture with streptavidin-coated magnetic beads
in the presence of Triton X-100 (0.1%), followed by the
described washing procedure and elution using the 2,5-dihy-
droxybenzoic acid matrix (21).
Photolabeling experiments with Pen(Lys48Bz) (2864.1) and
DMPG vesicles led to three pairs of signals: one pair (3092.3
and +16) corresponding to photoadduct with the myristoyl and
its oxidized form (+ 228 and + 244, respectively); another pair
(2922.1 and +16) that was assigned to the dehydrated form of
glycerol adducts formed during the basic treatment, and the third
pair corresponding to the starting peptide and its oxidized form.
Similar behavior and photoadducts were observed with
Pen(Lys48TMD). Besides the oxidized form of Pen(Lys48TMD),
previously observed in ammonium acetate buffer, two pairs of
signals were detected, corresponding to mass increases of 228
and 72, which were assigned to photoadducts of myristoyl and
oxidized/dehydrated glycerol, respectively. Finally, only de-
graded (low-molecular weights) forms of Pen(Lys48DTK) were
detected after irradiation with DMPG vesicles (Figure 3).
Effect on the Organization of the Phospholipid Bilayer.
DSC is used to study peptide/lipid interaction by monitoring
the ability of peptides to interact and perturb the lipid acyl chain
packing (Figure 1). A lipid bilayer composed of a single
phospholipid, in the presence of water excess, can undergo two
phase transitions upon heating: pretransition and main phase
DISCUSSION
All experiments were performed in air-saturated water, in
order to mimic the cellular environment. It has to be noted that,
in the experiments described herein, 10 nmol of peptide was
irradiated in the presence of about 90 nmol of oxygen (air-
saturated water). Oxygen or water must be the first source(s),
which might prevent the bimolecular photo-cross-linking of the
photoactivated probes. The lifetime of the benzophenone triplet
is 1.5 µs in air-saturated water, and much longer (20 µs) in
degassed water (31). Oxygen efficiently quenches the benzophe-
none triplet by energy transfer, leading to the reactive singlet
oxygen species with a short lifetime 4.2 µs in water (32), and
its deactivation occurs by light emission and electronic-to-
vibrational energy transfer and by intermolecular interaction with
specific amino acids such as Phe or Trp, which are both present
in penetratin (33). The benzophenone triplet can also be
quenched by intra- or intermolecular interactions, by charge
transfer, hydrogen abstraction, or light emission. However, all
these reactions are fast and reversible, with the exception of
the hydrogen abstraction step. The concentrations of the two
active forms, i.e., the ketone triplet and the singlet oxygen, must
be low under the irradiation conditions used herein, since the
consumption of the starting peptide is slow (total disappearance
occurs in over 2 h under permanent irradiation). For the dithienyl
ketone triplet in air-saturated water, a large proportion of
oxidized forms was observed showing that the singlet oxygen
concentration must be much higher than for the benzophenone
triplet. The carbene generated from trifluoromethylaryldiazirines
has a very short lifetime (t1/2 ≈ ns) (34) because it reacts rapidly
and irreversibly with oxygen. However, this competitive reaction
with oxygen does not totally prevent the hydrogen abstraction,
which can either occur intra- or intermolecularly.
With the benzophenone-substituted penetratin analogue, we
have shown that the gradual decrease as a function of time up
to the total disappearance of the starting peptide is slow. The
intermolecular self-reaction is unlikely since no dimer was
observed by MALDI-TOF MS (data not shown). The intramo-
lecular reactions (hydrogen abstraction and radical combination)
leading to cyclic peptides cannot be excluded with the techniques
used. Photo-cross-linking with other species has already been
reported for hydrophobic molecules (peptides, fatty acids,

356 Bioconjugate Chem., Vol. 21, No. 2, 2010
Jiao et al.
Figure 2. Quantification of remaining Pen(Lys48Bz) by MALDI-TOF mass spectrometry, after different irradiation times (30, 60, and 120 min
irradiation) in glycerol. Insert top right, quantification of the areas of B vs A. A corresponds to 2 pmol of the peptide with four hydrogenated
glycines added after the irradiation as an internal standard, the irradiation being performed with the same peptide bearing four deuterated glycines;
m/z correspond to peaks B. For the capture by streptavidin-coated magnetic beads and the detection by MALFDI-TOF mass spectrometry as
described in the experimental section.
Table 1. Thermodynamic Parameters Obtained by DSC for the
Interaction of Penetratin and Its Photoactivatable Analogues with
DMPG MLVs at P/L of 1/25
phospholipid photoadducts were detected, even if the lateral
diffusion in the disordered lamellar phase is around 10-7 > TD
< 5.10-9 cm2.s-1 (40) with fast phospholipid motion (10-12 to
10-7 s) and fast exchange between free and bound peptide (with
a dissociation constant in the micromolar range). The combina-
tion reaction occurs at time scales (10-6 s) slower than most
phospholipid motion. Thus, the detection of photo-cross-linking
with DMPG vesicles indicates the following: (i) cooperative
diffusion of the penetratin photoactivatable analogues and the
surrounding anionic phospholipids and (ii) restricted motion of
the anionic phospholipids.
pretransition
main transition
∆H (kcal/mol) Tp (° C) ∆H (kcal/mol) Tm (° C)
lipid alone
penetratin
Pen(LysBz)
Pen(LysDTK)
Pen(LysTMD)
0.4
-
-
0.2
-
9.6
-
6.2
3.8
4.6
6
22.3
22.1
22.3
22.1
22.5
a
-
8.8
-
8
a - the transition was abolished.
In a polar microenvironment such as glycerol, the triplet
radical state can react by energy transfer like a photosensi-
tizer, which, in the presence of oxygen, leads to the formation
of singlet state oxygen. Finally, in a lipophilic medium the
triplet radical might react with compounds possessing carbon-
bound removable hydrogen atoms, which function as a
quencher, as long as the intrinsic movements of the two
molecules are slow enough to allow radical combination
(DMPG). The cross-linking yield with trifluoromethylaryl-
diazirines, Pen(Lys48TMD), was also relatively low due to
a competitive reaction, the carbene leading to a fast irrevers-
ible reaction with oxygen. The oxidation and ꢀ-elimination
of the polar headgroups by basic hydrolysis prevent the
assignment of the exact structures of the photoadducts.
Finally, the triplet radical of dithienylketone, Pen(Lys48DTK),
must act as a photosensitizer, regardless of the medium, since
only oxidized and degraded species (lower molecular mass
species) were found.
phospholipids, glycolipids) soluble/inserted into a lipophilic
phase, such as micelles or vesicles (8, 14, 35-38). A pioneer
work has been reported on the labeling of “lipid mimics”, i.e.,
sodium dodecylsulfate (SDS) micelles, with 10 mM benzophe-
none, using a 1/1 molar ratio (35). The very high peptide
concentration and peptide/lipid ratio used in this study allowed
the observation of the radical combination step, even though
SDS exchange between micelles and free components and
molecular motion (τc between 10-10 and 10-12 s) is a fast process
(t1/2 ) 10-7 s) (39). Because the biological effects of membrane
active peptides are generally observed at micromolar concentra-
tion and P/L ratios of ∼1/50, cross-linking experiments pre-
sented here have been performed at micromolar concentrations.
No efficient cross-linking was observed between the three
tested photoactivatable peptides, at micromolar concentration
in the presence of SDS micelles. These results suggest that a
high concentration of photoreactive species is required to
compensate the fast dynamics of supramolecular assemblage
of SDS in micelles.
Thus, among the three photoprobes tested herein, only
benzophenone (Bz) and trifluoromethylaryldiazirine (TMD) led
to covalent bond formation with lipids, which were identifiable
However, using DMPG vesicles and also micromolar concen-
trations of photoactivatable Pen(Lys48Bz) or Pen(Lys48TMD),

Lipid Photo-Cross-Linking with Penetratin Analogues
Bioconjugate Chem., Vol. 21, No. 2, 2010 357
Figure 3. MALDI-TOF mass spectra of Pen(Lys48Bz) (A), Pen(Lys48DTK) (B), and Pen(Lys48TMD) (C) obtained after irradiation in water, glycerol,
and DMPG vesicles. Pen(Lys48Bz), Pen(Lys48DTK), or Pen(Lys48TMD) dissolved at micromolar concentration in water or glycerol (500 µL) was
irradiated on ice for 1 h at 365 nm. Then, 1 µL of the irradiated sample was deposited on the sample holder for MALDI-TOF mass spectrometry analysis.
DMPG (1 mg/mL) was mixed with Pen(Lys48Bz), Pen(Lys48DTK), or Pen(Lys48TMD) at 1/25 peptide/lipid ratio. Theses samples were irradiated on ice
for 1 h, at the same conditions as above, followed by an overnight basic hydrolysis. The hydrolyzed sample containing 0.1% of Triton was incubated with
streptavidin-coated magnetic beads (100 µg, Dynabeads M280, Dynal). The same procedure was used for the experiments performed in SDS, except that
no Triton was added during the incubation step. After different steps of washings, the photoadducts were eluted from the magnetic beads with the 2,5-
dihydroxybenzoic acid (DHB) matrix. After 10 min incubation, 1 µL of bead-free supernatant was deposited on the sample holder for MALDI-TOF mass
spectrometry analysis.

358 Bioconjugate Chem., Vol. 21, No. 2, 2010
Jiao et al.
de Kruijff, B., and Killian, J. A. (2004) Photo-crosslinking
analysis of preferential interactions between a transmembrane
peptide and matching lipids. Biochemistry 43, 4482–4489.
(9) Maxfield, F. R. (2002) Plasma membrane microdomains. Curr.
Opin. Cell Biol. 14, 483–487.
(10) Engelman, D. M. (2005) Membranes are more mosaic than
fluid. Nature 438, 578–580.
by MALDI-TOF mass spectrometry. Benzophenone (Bz) ap-
peared to be the most appropriated photoprobe compared to
TMD, since Bz leads to similar phase transition effects to those
of penetratin, while TMD leads to great changes in the
cooperativity of the transition, suggesting that TMD probably
inserts deeper into the fatty acid chain region.
(11) Lenne1, P.-F., Wawrezinieck, L., Fabien Conchonaud, F.,
Wurtz, O., Boned, A., Guo, X.-J., Rigneault, H., He, H.-T., and
Marguet, D. (2006) Dynamic molecular confinement in the
plasma membrane by microdomains and the cytoskeleton
meshwork. EMBO J. 25, 3245–3256.
(12) Lagerholm, B. C., Weinreb, G. E., Jacobson, K., and
Thompson, N. L. (2005) Detecting microdomains in intact cell
membranes. Ann. ReV. Phys. Chem. 56, 309–333.
(13) Chachaty, C., Rainteau, D., Tessier, C., Quinn, P. J., and Wolf,
C. (2005) Building up of the liquid-ordered phase formed by
sphingomyelin and cholesterol. Biophys. J. 88, 4032–4044.
(14) Bino, J. B., Kumar, E. R., and Lala, A. K. (2000) Depth-
dependent analysis of membranes using benzophenone-based
phospholipids. Biophys. Chem. 87, 37–42.
(15) Coates Barclay, L. R., Baskin, K. A., Locke, S. J., and
Schaefer, T. D. (1987) Benzophenone-photosensitized autoxi-
dation of linoleate in solution and sodium dodecyl sulfate
micelles. Can. J. Chem. 65, 2529–2540.
(16) Zhu, W. L., and Shin, S. Y. (2009) Antimicrobial and cytolytic
activities and plausible mode of bactericidal action of the cell
penetrating peptide penetratin and its Lys-linked two-stranded
peptide. Chem. Biol. Drug Des. 73, 209–15.
(17) Palm, C., Netzereab, S., and Ha¨llbrink, M. (2006) Quantita-
tively determined uptake of cell-penetrating peptides in non-
mammalian cells with an evaluation of degradation and antimi-
crobial effects. Peptides 27, 1710–6.
(18) Balayssac, S., Burlina, F., Convert, O., Bolbach, G., Chassaing,
G., and Lequin, O. (2006) Comparison of penetratin and other
homeodomain-derived cell-penetrating peptides: interaction in
a membrane-mimicking environment and cellular uptake ef-
ficiency. Biochemistry 45, 1408–1420.
(19) Christiaens, B., Symoens, S., Vanderheyden, S., Engelborghs,
Y., Joliot, A., Prochiantz, A., Vandekerckhove, J., Rosseneu, M.,
and Vanloo, B. (2002) Tryptophan fluorescence study of the
interaction of penetratin peptides with model membranes. Eur.
J. Biochem. 269, 2918–2926.
CONCLUSION
Herein, the photo-cross-linking of three photoactivatable
penetratin peptides in glycerol, SDS, and lipid (DMPG) was
evaluated. The dynamics of both SDS and the SDS/penetratin
assemblies and the micromolar concentration of photoacti-
vatable penetratin analogues used (in the range of biologically
active concentrations) prevented peptide cross-linking to SDS.
Regarding glycerol and lipids, even though photo-cross-
linking occurred in both cases, the direct MS analysis of the
photoadducts formed upon irradiation of the three different
photoactivatable penetratin analogues was only achieved in
the first one. The characterization of the photoadducts formed
after lipid irradiation required basic hydrolysis of the lipid
before extraction; otherwise, the capture with streptavidin-
coated magnetic beads was ineffective. Bilayers of DMPG
were photo-cross-linked with penetratin analogues, containing
either benzophenone or trifluoromethylaryl diazirines but not
phenylketone. With both photoactivatable penetratin ana-
logues, MALDI-TOF analysis of the photoadducts formed
after irradiation of the peptides with DMPG allowed unam-
biguous identification of the covalent bond formed with this
lipid. Therefore, these analogues can be used to analyze the
specific lipid(s) recruitment by penetratin, which represents
an essential step in the understanding of the translocation
mechanism of this cationic cell-penetrating peptide.
ACKNOWLEDGMENT
This work was supported by the Association Nationale pour
la Recherche (ANR-Prob DOM). We thank the Pr. Gilbert
Kirsch and his co-workers from the “Laboratoire d’Inge´nierie
Mole´culaire et Biochimie Pharmacologique” Metz, France, for
the generous gift of 5-carboxy-5′-methyl-2,2′-dithienylketone.
(20) Christiaens, B., Grooten, J., Reusens, M., Joliot, A., Goethals,
M., Vandekerckhove, J., Prochiantz, A., and Rosseneu, M. (2004)
Membrane interaction and cellular internalization of penetratin
peptides. Eur. J. Biochem. 271, 1187–1197.
(21) Burlina, F., Sagan, S., Bolbach, G., and Chassaing, G. (2005)
Quantification of the cellular uptake of cell-penetrating peptides
by MALDI-TOF mass spectrometry. Angew. Chem., Int. Ed. 44,
4244–4247.
LITERATURE CITED
(1) Park, E. J., Carroll, G. T., Turro, N. J., and Koberstein, J. T.
(2009) Shedding light on surfaces using photons to transform
and pattern material surfaces. Soft Matter 5, 36–50.
(2) Dorma´n, G., and Prestwich, G. D. (2000) Using photolabile
ligands in drug discovery and development. TIBTECH 18, 64–
76.
(22) Aussedat, B., Sagan, S., Chassaing, G., Bolbach, G., and Burlina,
F. (2006) Quantification of the efficiency of cargo delivery by
peptidic and pseudo-peptidic trojan carriers using MALDI-TOF
mass spectrometry. Biochem. Biophys. Acta 1758, 375–383.
(23) Delaroche, D., Aussedat, B., Aubry, S., Chassaing, G., Burlina,
F., Clodic, G., Bolbach, G., Lavielle, S., and Sagan, S. (2007)
Tracking a new cell-penetrating peptide, (RW)9, through an
enzyme-stable mass spectrometry reporter. TAG. Anal. Chem.
79, 1932–1938.
(24) Magzoub, M., Eriksson, L. E., and Gra¨slund, A. (2002)
Conformational 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.
(25) Binder, H., and Lindblom, G. (2003) Charge-dependent
translocation of the Trojan peptide penetratin across lipid
membranes. Biophys. J. 85, 982–95.
(26) Brunner, J., Senn, H., and Richards, F. M. (1980) 3-Triflu-
oromethyl-3-phenyldiazirine a new carbene generating group for
photolabeling reagents. J. Biol. Chem. 255, 3313–3318.
(3) Brunner, J. (1996) Use of photocrosslinkers in cell biology.
Trends Cell Biol. 6, 154–157.
(4) Kotzyba-Hibert, F., Kapfer, I., and Goeldner, M. (1995) Recent
trends in photoaffinity labeling. Angew. Chem., Int. Ed. 34, 1296–
1312.
(5) Fillion, D., Derae¨t, M., Holleran, B. J., and Escher, E. (2005)
Stereospecific synthesis of a carbene-generating angiotensin ii
analogue for comparative photoaffinity labeling: improved
incorporation and absence of methionine selectivity. J. Med.
Chem. 48, 3980–3990.
(6) Lequin, O., Bolbach, G., Fabrice, F., Convert, O., Girault-
Lagrange, S., Chassaing, G., Lavielle, S., and Sagan, S. (2002)
Involvement of the second extracellular loop (E2) of the
neurokinin-1 receptor in the binding of substance P: Photoaffinity
labeling and modeling studies. J. Biol. Chem. 277, 22386–22394.
(7) Jullian, V., Courtois, F., Bolbach, G., and Chassaing, G. (2003)
Carbon-carbon bond ligation between cyclodextrin and peptide
by photo-irradiation. Tetrahedron Lett. 44, 6437–6440.
(8) Ridder, A. N. J. A., Spelbrink, R. E. J., Demmers, J. A. A.,
Rijkers, D. T. S., Liskamp, R. M., Brunner, J., Heck, A. J. R.,

Lipid Photo-Cross-Linking with Penetratin Analogues
Bioconjugate Chem., Vol. 21, No. 2, 2010 359
(27) Nassal, M. (1983) 4-(1-Azi- 2,2,2-trifluoroethyl)benzoic
acid a highly photolabile carbene generating label readly
fixable to biochemical agents. Liebigs Ann. Chem. 1510–1523.
(28) Derossi, D., Joliot, A. J., Chassaing, G., and Prochiantz, A.
(1994) The third helix of the Antennapedia homeodomain
translocates through biological membranes. J. Biol. Chem. 269,
10444–10450.
(29) McElhaney, R. N. (1986) Differential scanning calorimetric
studies of lipid-protein interactions in model membrane systems.
Biochim. Biophys. Acta 864, 361–421.
implications for the fusion mechanism. J. Gen. Virol. 76, 1541–
1556.
(35) Breslow, R., Kitabatake, S., and Rothbard, J. (1978) Photo-
reactions of charged benzophenone with amphiphiles in micelles
and multicomponent aggregates as conformational probes. J. Am.
Chem. Soc. 100, 8156–8160.
(36) Delfino, J. M., Schreiber, S. L., and Richards, F. M. (1993)
Design, synthesis, and properties of a photoactivatable membrane-
spanning phospholipidic probe. J. Am. Chem. Soc. 115, 3458–
3474.
(37) Lala, A. K., and Kumar, E. R. (1993) Orientation of the
benzophenone group at various depths in bilayers. J. Am. Chem.
Soc. 115, 3982–3988.
(38) Wendeler, M., Hoernschemeyer1, J., Hoffmann, D., Kolter,
T., Schwarzmann, G., and Sandhoff, K. (2004) Photoaffinity
labeling of the human GM2-activator protein mechanistic insight
into ganglioside GM2 degradation. Eur. J. Biochem. 271, 614–
627.
(30) Alves, I. D., Goasdoue´, N., Correia, I., Aubry, S., Galanth,
C., Sagan, S., Lavielle, S., and Chassaing, G. (2008) Membrane
interaction and perturbation mechanisms induced by two cationic
cell penetrating peptides with distinct charge distribution. Bio-
chim. Biophys. Acta 1780, 948–959.
(31) Canonica, S., Hellrung, B., and Wirz, J. (2000) Oxidation of
Phenols by Triplet Aromatic Ketones in Aqueous Solution. J.
Phys. Chem. A 104, 1226–1232.
(39) Barker, C., Saul, D., Tiddy, G., Wheeler, B., and Willis, E.
(1974) Phase structure, nuclear magnetic resonance and. rheo-
logical properties of viscoelastic sodium. dodecyl sulphate and
trimethylammonium bromide mixtures. J. Chem. Soc., Faraday
Trans 1 70, 154–162.
(40) Jain, M. K., and White, H. B. (1977) Long-range order in
biomembranes. AdV. Lipid Res. 15, 1–60.
(32) Rodgers, M. A. J. (1983) Solvent-induced deactivation of
singlet oxygen: additivity relationships in nonaromatic solvents.
J. Am. Chem. Soc. 105, 6201–6205.
(33) Harada, Y., Suzuki, K., Hashimoto, M., Tsukagoshi, K. and
H., and Kimoto, H. (2009) Chemiluminescence from singlet
oxygen that was detected at two wavelengths and effects of
biomolecules on it. Talanta 77, 1223–1227.
(34) Gaudin, Y., Ruigrok, R. W. H., and Brunner, J. (1995) Low-
pH induced conformational changes in viral fusion proteins:
BC900466Q