Photoactivatable phospholipids bearing tetrafluorophenylazido chromophores exhibit unprecedented protonation-state-dependent 19F NMR signals

Article abstract of DOI:10.1021/ol201593k

Phospholipids bearing tetrafluorophenylazido chromophores were synthesized with perfectly conserved amphiphilicity and photochemical activity. Interestingly however, those phospholipids harboring the amine-linked chromophores exhibited unusual ¹⁹

Full text of DOI:10.1021/ol201593k

ORGANIC
LETTERS
2011
Vol. 13, No. 16
4248–4251
Photoactivatable Phospholipids
Bearing Tetrafluorophenylazido
Chromophores Exhibit
Unprecedented Protonation-State
-Dependent 19F NMR Signals
Yi Xia,†,‡ Fanqi Qu,‡ Alain Maggiani,† Kheya Sengupta,† Cheng Liu,‡ and
Ling Peng*,†
State Key Laboratory of Virology, College of Chemistry and Molecular Sciences,
Wuhan University, Wuhan 430072, P.R. China, and Centre Interdisciplinaire de
Nanoscience de Marseille, CNRS UPR 3118, 13288 Marseille, France
ling.peng@univmed.fr
Received June 14, 2011
ABSTRACT
Phospholipids bearing tetrafluorophenylazido chromophores were synthesized with perfectly conserved amphiphilicity and photochemical
activity. Interestingly however, those phospholipids harboring the amine-linked chromophores exhibited unusual 19F NMR signals which
depended on the protonation state of the lipid headgroup. These probes may serve as powerful tools for studying various pH-dependent events in
biomembranes.
Biomembranes are of considerable importance not only
due to their function as the cell boundary regulating the
exchange with and protection from the outside environ-
ment but also due to their multiple roles in intra- and
extracellular processes.1 However, studying these biomem-
branes is extremely challenging due to their heterogeneous
mosaic composition and limited solubility. Photoactivata-
ble phospholipids are useful tools to investigate biomem-
branes via a photolabeling approach.2,3 In our effort to
develop stable and efficient photoactivatable phospho-
lipidic probes, we are particularly interested in using
tetrafluorophenylazide moieties as photoactive chromo-
phores.4,5 This is due to much more efficient photolabeling
yielded by the fluorinated aryl azides compared to their
nonfluorinated counterparts thanks to considerable stabi-
lization of the reactive arylnitrene intermediates by the
electronegative fluorine atoms and the resulting longer
half-life.6 Moreover, the presence of fluorine atoms7,8
provides a unique advantage in 19F NMR based investiga-
tions of biological processes thanks to the absence of
fluorinated compounds in most biological samples which
could otherwise give rise to interfering background signals.
† CINaM CNRS UPR 3118.
(4) Peng, Q.; Xia, Y.; Qu, F.; Wu, X.; Campese, D.; Peng, L.
Tetrahedron Lett. 2005, 46, 5893–5897.
(5) Huang, F.; Qu, F.; Peng, Q.; Xia, Y.; Peng, L. J. Fluorine Chem.
‡ Wuhan University.
(1) van Meer, G.; Voelker, D. R.; Feigenson, G. W. Nat. Rev. Mol.
2005, 126, 739–743.
Cell Biol. 2008, 9, 112–124.
(6) Gritsan, N. P.; Platz, M. S. Chem. Rev. 2006, 106, 3844–3867.
(7) Reichenbacher, K.; Suss, H. I.; Hulliger, J. Chem. Soc. Rev. 2005,
34, 22–30.
(2) Peng, L.; Alcaraz, M. L.; Klotz, P.; Kotzyba-Hibert, F.; Goeldner,
M. FEBS Lett. 1994, 346, 127–131.
(3) Kotzyba-Hibert, F.; Kapfer, I.; Goeldner, M. Angew. Chem., Int.
€
Ed. Engl. 1995, 34, 1296–1312.
(8) Muller, K.; Faeh, C.; Diederich, F. Science 2007, 317, 1881–1886.
r
10.1021/ol201593k
Published on Web 07/18/2011
2011 American Chemical Society

The high sensitivity of 19F NMR results additionally from
the 100% natural abundance of 19F as well as the sensitive
resonance signals distributed over a wide spectral width
upon even minor chemical changes.9
Scheme 2. Synthesis of Probes 3 (A) and 4 (B)
Scheme 1. Photoactivatable Phospholipidic Probes Containing
the Tetrafluorophenylazido Group
In our previous work, we synthesized two photoactiva-
table phospholipidic probes 1 and 2 containing the photo-
active tetrafluorophenylazido group at the polar head and
in the fatty acid chain, respectively (Scheme 1).4 However,
probe 1 has the chromophore coupled to the phosphati-
dylethanolamine via amide linkage, leading to the abolish-
ment of the positive chargeofthe phospholipid head amine
group at physiological pH. This may alter the related
structure and functions of phospholipids when incorpo-
rated within biomembranes. Meanwhile probe 2, which
has the tetrafluorophenylazido group attached to the fatty
acyl chain via an ester linkage, may be hydrolytically or
enzymatically labile and thus limit its potential applica-
tions. We therefore wanted to introduce the tetrafluoro-
phenylazido group to the phospholipids via a more reliable
amine and ether linkage at the polar head and in the fatty
acid chain respectively (probes 3 and 4 in Scheme 1). The
amine linkage between the chromophore and the amine
head of the phospholipids was expected to ensure that the
phospholipid heads of the probes 3aꢀc retained their
protonation ability at physiological pH, while the ether
linkage between the tetrafluorophenylazido group and the
fatty acid was expected to make probe 4 resistant to
chemical and enzymatic hydrolysis. Here we report the
synthesis and characterization of these two probes. Inter-
estingly, probe 3 exhibits unusually broad 19F NMR
signals arising from the different protonation states of
the adjacent amine function at the phospholipid head-
group. This finding could be useful when studying the
biological processes involving pH variation within bio-
membranes that are important and yet difficult to investi-
gate due to the absence of suitable and sensitive methods.
Probe 3 was synthesized by the reductive amination of
4-azidotetrafluorobenzaldehyde with the corresponding
phosphatidylethanolamine (Scheme 2A) using sodium tria-
cetoxyborohydride [NaBH(OAc)3].10,11 In addition to the
expected product 3, we also obtained two side products
arising from the direct reduction of 4-azidotetrafluoro-
benzaldehyde and the double reductive amination of 4-azi-
dotetrafluorobenzaldehyde, respectively (Scheme S1). In
order to diminish these side reactions, we reduced the excess
amount of 4-azidotetrafluorobenzaldehyde and shortened
the reaction time. To further improve the reaction, we chose
CHCl3 as the solvent for its good solubility of phospholipid
compounds. Altogether, these efforts allowed us to obtain
probes 3aꢀc with good and satisfactory yields.
Synthesis of probe 4 was then achieved by conjugating
the tetrafluorophenylazido-containing fatty acid surrogate
5 with lyso-phosphocholine (lyso-PC) (Scheme 2B). Com-
pound 6, the ester form of 5, was prepared by coupling 7
with 4-azidotetrafluorobenzyl alcohol, with 7 being ob-
tained conveniently using a one-step procedure developed
by ourgroup (Scheme S2). Weobtained 6 in only moderate
yield, with the best yield being attained using the AgBF4/
Ag2CO3 system.12 This was mainly due to the low reactiv-
ity of 7 caused by its long alkyl chain, because the shorter
alkyl iodides such as CH3I and CH3CH2I could give
(9) Danielson, M. A.; Falke, J. J. Annu. Rev. Biophys. Biomol. Struct.
1996, 25, 163–195.
(10) Abdel-Magid, A. F.; Carson, K. G.; Harris, B. D.; Maryanoff,
C. A.; Shah, R. D. J. Org. Chem. 1996, 61, 3849–3862.
(11) Chehade, K. A.; Kiegiel, K.; Isaacs, R. J.; Pickett, J. S.; Bowers,
K. E.; Fierke, C. A.; Andres, D. A.; Spielmann, H. P. J. Am. Chem. Soc.
2002, 124, 8206–8219.
(12) Bhatia, S. K.; Hajdu, J. J. Org. Chem. 1988, 53, 5034–5039.
Org. Lett., Vol. 13, No. 16, 2011
4249

satisfactory yields.12 The obtained 6 was then hydrolyzed to
give 5 (Scheme 2B). Coupling of 5 with lyso-PC led to the
desired probe 4. Despite numerous attempts to improve the
yield of 4 via various means such as performing the reaction
under strictly anhydrous conditions, or using different
activating or coupling reagents, it proved difficult to
achieve higher yields mainly due to the incomplete coupling
reaction but also the extremely difficult purification process
by performing flash chromatography on silica gel and the
notorious instability of the phospholipid compound.
addition, the observed isosbestic points indicate that the
photochemical reaction was a clear-cut photodecomposi-
tion process. Overall, these data indicate that our lipid
probes retain the photochemical properties of fluorinated
phenylazide and are potentially promising tools in photo-
labeling studies of biomembranes.
Figure 3. 19F NMR spectra of the photoactivatable phospholi-
pidic probes 3a (A), 3b (B), 3c (C), and 4 (D) were recorded in
CDCl3.15
We went on to analyze the 19F NMR of these probes
(Figure 3) as 19F NMR is an alternative and com-
plementary approach to photolabeling to investigate
biomembranes.14 To our surprise, probe 3 showed an
unusually broad 19F signal around ꢀ144 ppm
(Figure 3Aꢀ3C), in addition to the normal and sharp
singlet peak recorded at ꢀ152 ppm. This was completely
different from what we previously observed with lipid
probes containing a 4-azidotetrafluorophenyl chromo-
phore including probes 1, 2 (Figure S3), and 4
(Figure 3D), all of which show two sharp 19F NMR signals
around ꢀ153 and ꢀ144 ppm, corresponding to the two
pairs of fluorine atoms on the phenyl ring in the chromo-
phore. Similar to the phospholipids 1, 2, and 4, the lipid
probes 3aꢀc are well soluble in the NMR solvent and,
therefore, exclude the formation of micelles or vesicles,
which might be the origin of the observed unusual NMR
signals.
We know that the amine functionality can be easily
protonated in physiological media at a pH less than its
pKa value. Based on this knowledge, we hypothesized that
the free and protonated states of the amine function in
probe 3 may cause different chemical shifts of the adjacent
F atoms in the tetrafluorophenylazido chromophore; the
equilibrium between these two states could lead to the
merge of the two distinct 19F NMR signals into one
broader peak (Figure 3Aꢀ3C). In order to confirm this,
we performed 19F NMR analysis on a model compound in
its free amine 8 and protonated amine 80 states (for their
preparation see Supporting Information), rather than on 3
directly. This is because the phospholipid probe 3 is
relatively fragile in its protonated state at low pH during
the 19F NMR experimental process. With the aid of this
model compound, we indeed observed that the 19F nuclei
in the chromophore adjacent to the amine function had
two sharp and distinct chemical shifts at ꢀ145 and ꢀ139
Figure 1. Pressureꢀarea isotherms of surface monolayers of the
synthesized probes 3b (A) and 4 (B).
All the synthesized probes gave excellent NMR, IR, UV,
and HRMS analytical data. We then wished to test
whethertheyhad conservedtheamphiphiliccharacteristics
of natural phospholipids. All the probes were poorly
soluble in water; however, their solubility could be con-
siderably increasedinthe presenceof the detergent, sodium
dodecylsulfonate (Figure S1). In addition, these probes
could form a Langmuir monolayer film at the air/water
interface.13 Their pressureꢀarea isotherm curves (Figure 1)
are strikingly similar and resemble that obtained with the
natural phospholipid, 1,2-dioleoyl-sn-glycero-3-phospho-
choline (Figure S2). This suggests that the differences in
their structure do not influence the area occupied by each
molecule or the intermolecular interactions. Collectively,
these data confirm that probes 3 and 4 are, like natural
phospholipids, amphiphilic.
We next studied the photochemical properties of these
phospholipid probes. Both 3 and 4 were stable in the dark
and underwent a rapid photochemical reaction (Figure 2).
Irradiation of these probes at 300 nm quickly led to the
disappearance of the maximal absorption band of the
tetrafluorophenylazide chromophore (Figure 2). In
(13) Maget-Dana, R. Biochim. Biophys. Acta 1999, 1462, 109–140.
(14) Ulrich, A. S. Prog. Nucl. Magn. Reson. Spectrosc. 2005, 46, 1–21.
(15) Two or three drops of CD3OD were added to 300ꢀ500 μL of
CDCl3 in order to improve NMR resolution.
Figure 2. UV spectral recording of the photochemical reaction
of 3b (A) and 4 (B) upon irradiation at 300 nm at 20 °C.
4250
Org. Lett., Vol. 13, No. 16, 2011

Cell adhesion, biomembrane fission and fusion, lipidꢀpro-
tein binding, etc. depend critically on intracellular pH and
the protonation state of the phospholipid headgroup.19,20
The photoactivatable phospholipid probe 3 developed here
could constitute a useful means to study these processes by
19F NMR in addition to the photolabeling approach.
Moreover, 19F NMR has the unique advantage of being
highly sensitive and informative for the study of biological
systems using fluorinated probes21 because of the absence of
natural fluorinated compounds in most biological systems,
the high natural abundance of fluorine, and the wide range
of chemical shifts sensitive to the chemical environment.
In conclusion, we have synthesized novel photoactiva-
table phospholipidic probes containing tetrafluoropheny-
lazido groups either at the polar head via an amine bridge
or in the fatty acid chain via an ether linkage. The lipid-like
amphiphilic characteristics and excellent photochemical
properties of these probes forecast their potential applica-
tion in photolabeling studies of biomembranes. Further-
more, when linked to the amine function at the phospho-
lipid headgroup, the tetrafluorophenylazide chromophore
exhibited protonation state-dependent 19F NMR signals,
which could prove useful in investigating pH-dependent
membrane processes. These are often difficult to study but
are nevertheless involved in important phenomena includ-
ing lipidꢀprotein interactions, membrane fission/fusion,
drug delivery, etc.
Figure 4. 19F NMR spectral recording of 8, 80, and their mixture
at a ratio of 1/1 were recorded in CDCl3.
ppm in free amine 8 and protonated amine 80, respectively
(Figure 4). By mixing 8 with 80 in a 1:1 ratio,16 the 19F
NMR signals at ꢀ145 ppm (8) and ꢀ139 ppm (80) dis-
appeared and coalesced into a single, exchange broadened
NMR signal at ꢀ142 ppm (Figure 4). This clearly indicates
that there is chemical exchange between the two amine
forms 8 and 80, with an exchange rate in the order of the
chemical shift separation (expressed in Hz). Indeed, slow
chemical exchange on the NMR time scale would have led
to the observation of two distinct 19F signals, one due to
the free (ꢀ145 ppm) and the other to the protonated amine
(ꢀ139 ppm). In contrast, fast chemical exchange would
lead to the observation of a single line, the chemical shift of
which would be the average of the chemical shifts of the
two forms (i.e., around ꢀ142 ppm for a 1:1 molar ratio).
Only intermediate chemical exchange rates can yield ex-
change broadened signals similar to those depicted in
Figures 3 and 4.
Acknowledgment. We are grateful for the financial
support from National Mega Project on Major Drug
Development (2009ZX09301-014), ANR “ProKrebs”pro-
ject (07-PCVI-0028), Wuhan University, CNRS, and
French Embassy in China for the short-term fellowship
ꢀ
to support Y.X. We thank Drs. Stephane Viel (Aix-
Marseille University), Fabio Ziarelli (Spectropole,
^
Marseille), Han Chen (Harvard Medical School), and
Mr. Yang Wang (CINaM UPR 3118) for helpful discussions
and critical reading of the manuscript.
It is known that the electron-withdrawing character of
the fluorine atom can affect the pKa values of the corre-
sponding benzylic amines, as shown by the pH back-
titration of the compound 8 and benzyl amine (Figure
S4). This can be interesting as the fluorinated benzylic
amines may also serve as 19F NMR pH indicators for
an investigation on pH dependent biological events.17,18
Supporting Information Available. Experimental pro-
cedures and spectroscopic data for all new compounds
as well as Schemes S1ꢀS3 and Figures S1ꢀS4. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(16) Due to the poor solubility of probe 3 and compound 8 in water,
we could not undertake a reliable pH titration process of these probes
using 19F NMR.
(19) Young, B. P.; Shin, J. J. H.; Orij, R.; Chao, J. T.; Li, S. C.; Guan,
X. L.; Khong, A.; Jan, E.; Wenk, M. R.; Prinz, W. A.; Smits, G. J.;
Loewen, C. J. R. Science 2010, 329, 1085–1088.
(17) He, S.; Mason, R. P.; Hunjan, S.; Mehta, V. D.; Arora, V.;
Katipally, R.; Kulkarni, P. V.; Antich, P. P. Bioorg. Med. Chem. 1998, 6,
1631–1639.
(20) Lahdesmaki, K.; Ollila, O. H.; Koivuniemi, A.; Kovanen, P. T.;
Hyvonen, M. T. Biochim. Biophys. Acta 2010, 1798, 938–46.
ꢀ
(21) Laurent, S.; Chen, H.; Bedu, S.; Ziarelli, F.; Peng, L.; Zhang, C.-
(18) Deutsch, C. J.; Taylort, J. S. Biophys. J. 1989, 55, 799–804.
C. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 9907–9912.
Org. Lett., Vol. 13, No. 16, 2011
4251