Rationalizing the F...S interaction discovered within a tetrafluorophenylazido-containing bola-phospholipid

Article abstract of DOI:10.1039/c2cc18147a

A bola-lipid bearing tetrafluorophenylazido chromophore in the diacyl chain displayed puzzling ¹⁹F NMR, leading to the evidence and rationalization of a F...S weak interaction that is important for altering molecular structures and imposing nov

Full text of DOI:10.1039/c2cc18147a

View Article Online / Journal Homepage / Table of Contents for this issue
Chemical Communications
www.rsc.org/chemcomm
Volume 48 | Number 36 | 7 May 2012 | Pages 4261–4376
ISSN 1359-7345
COMMUNICATION
Ling Peng et al.
Rationalizing the F···S interaction discovered within a tetrafluorophenylazido-containing bola-phospholipid

Dynamic ArticVleiewLiAnrktiscle Online
ChemComm
Cite this: Chem. Commun., 2012, 48, 4284–4286
www.rsc.org/chemcomm
COMMUNICATION
Rationalizing the FÁ Á ÁS interaction discovered within a
tetrafluorophenylazido-containing bola-phospholipidw
c
a
d
e
e
Yi Xia,ab Stephane Viel, Yang Wang, Fabio Ziarelli, Erik Laurini, Paola Posocco,
´
Maurizio Fermeglia,e Fanqi Qu,b Sabrina Pricle and Ling Peng*a
Received 29th December 2011, Accepted 27th January 2012
DOI: 10.1039/c2cc18147a
A bola-lipid bearing tetrafluorophenylazido chromophore in the
diacyl chain displayed puzzling 19F NMR, leading to the
evidence and rationalization of a FÁ Á ÁS weak interaction that
is important for altering molecular structures and imposing novel
and special properties on fluorinated compounds.
tetrafluorophenylazido chromophore is introduced through a
thioacetal bridge into the middle of the diacyl acid chain. The
weak FÁÁÁS interaction disclosed here is of significance with respect
to altering molecular structures and imposing novel and special
properties on fluorinated compounds. We present our work below
aiming at rationalizing this remarkable FÁÁÁS interaction.
The bola-phospholipid 1 was synthesized using a reported
procedure (Scheme S1, ESIw).2 1 gave all satisfactory spectral
analysis except the 19F NMR spectral data, which displayed
surprisingly only one distinct resonance (at À152 ppm) (entry 1
in Table 1), in contrast to our previously synthesized probes3 or
any other reported compounds containing tetrafluorophenylazide
for which two groups of 19F resonances can be expected, each one
corresponding to one pair of fluorine atoms (in the ortho and meta
position with respect to the azido function, respectively).4 To clarify
this discrepancy, we recorded the 19F NMR spectrum of 2 (entry 2
in Table 1), the precursor for synthesizing 1 (Scheme S1, ESIw).
Similarly, only one sharp 19F signal at À151 ppm was observed.
Interestingly, by replacing the fatty acyl chain in 2 with two shorter
and smaller propyl groups, the 19F NMR spectrum of the resulting
compound 3 displayed a broad signal at around À135 ppm in
addition to the sharp signal at À151 ppm (entry 3 in Table 1).
These results suggest that the bulky alkyl chains contribute to the
unexpectedly abnormal 19F NMR spectrum. We further sub-
stituted the thioacetal moiety in 3 with an acetal function to obtain
the model compound 4 (entry 4 in Table 1). In this case, two sharp
19F NMR signals were detected corresponding to the two pairs of
fluorine atoms at the ortho and meta positions, as expected.
Consequently, these data imply that the S atoms in the thioacetal
group are also responsible for the unexpected disappearance of the
second 19F NMR signal expected for 1.
The introduction of fluorine atoms into organic molecules often
imparts them with novel and special physicochemical and bio-
logical properties which are not only the direct consequence of
the strong C–F bond, but also often originate from the weak
intra- and intermolecular interactions engendered by the
C–F bond. While the inductive effects of the C–F bond are
relatively well understood, the need for a better understanding of
the special properties of fluorinated compounds has led to an
increased interest in the non-covalent weak dipolar interactions
such as C–FÁÁÁp, C–FÁÁÁCQO, C–FÁÁÁH–C, C–FÁÁÁH–X
(X = N, O...) etc. and their specific influence on molecular
structure and properties.1 For example, the above-mentioned
weak F interactions are found in the Protein Data Bank (PDB)
which favors selective protein–ligand interactions.1a Meanwhile,
the generally weak non-covalent interactions brought about by
selective aromatic fluorine substitution have been shown to
increase the affinity of a molecule for a macromolecular recogni-
tion site.1d In this work, we report a FÁÁÁS weak interaction
discovered serendipitously during our investigation on the photo-
activatable bola-lipid probe
1 (Fig. 1) in which the
In light of the above findings, we went on to verify which
pair of fluorine atoms in the tetrafluorophenylazido group in 1
corresponded to the missing 19F NMR signal. We thus
examined the 19F NMR spectra of the model compounds
5 and 6 (entries 5 and 6 in Table 1). While 5 displayed a single
19F peak as expected, 6 surprisingly exhibited two abnormally
broad, chemically non-equivalent signals. This information
indicated that the unusual 19F NMR signal was due to the
two fluorine atoms in the ortho position with respect to the
thioacetal group. By coupling this finding with those obtained
from 4 (entry 4 in Table 1), we proposed that the two S atoms
interact with the adjacent F atoms, and thereby impede the
Fig. 1 The bola-phospholipidic probe 1.
a
´
Aix-Marseille Universite, CNRS CINaM UMR 7325, Marseille,
France. E-mail: ling.peng@univmed.fr; Tel: +33 4918 29154
b College of Chemistry and Molecular Sciences, Wuhan University, China
c
´
Aix-Marseille Universite, Institut de Chimie Radicalaire,
CNRS UMR 7273, Marseille, France
d
´
Aix-Marseille Universite, Spectropole, Marseille, France
e Molecular Simulation Engineering Laboratory, Trieste University, Italy
w Electronic supplementary information (ESI) available: Experimental
section. See DOI: 10.1039/c2cc18147a
c
4284 Chem. Commun., 2012, 48, 4284–4286
This journal is The Royal Society of Chemistry 2012

View Article Online
Table 1 19F NMR room temperature spectra of various compoundsa
Entry
Compound
19F NMR
1
Probe 1
2
Fig. 2 Temperature-dependent 19F NMR spectra of 6 (A) and 9 (B).
The spectra marked with the red dashed frames are those recorded
near the corresponding coalescence temperatures.
3
4
Steric hindrance was reduced in 7 by decreasing the distance
between F and S (via the cyclic thioacetal structure), and totally
suppressed in 8 by removing one of the two S-containing alkyl
chains. The 19F NMR spectra of 7 and 8 displayed only one
relatively sharp signal (entries 7 and 8 in Table 1), suggesting that
the steric constraint between S and F did contribute to the
reduction in rotational freedom observed for 1, 2, 3 and 6.
However, the steric hurdle was not the only contributing factor.
Indeed, the 19F NMR spectrum of 9, in which each divalent S
atom was replaced by a CH2 moiety of similar size (24.04 A3 and
24.91 A3, respectively), also showed a single relatively sharp signal,
as opposed to the two extremely broad ones observed for 6. In
other words, although the steric congestion in 6 and 9 is essentially
comparable, the 19F NMR spectra of each of the compounds are
strikingly different. These results clearly suggest the eventual
presence of an additional contributing factor impeding the rota-
tion in the form of an orbital interaction between F and S.
To estimate the contribution of the interaction between F and S,
we performed a series of temperature-dependent 19F NMR experi-
ments to investigate the conformational dynamic processes within
the molecule.5 We primarily focused on compound 6 rather than
probe 1 because of the much higher thermal stability of 6. The
spectra obtained within the temperature range from À50 1C to
100 1C are reported in Fig. 2A and Fig. S3A (ESIw). As can be
seen, fluorine atoms in 6 showed chemical non-equivalence at low
temperatures, resulting in the observation of two distinct
resonances, whereas only one resonance was observed at high
temperatures, with a coalescence temperature (TC) near 32 1C. This
value allowed a free energy of activation for rotation to be
estimated (12.8 Æ 0.2 kcal molÀ1). This result agreed well with
theoretical calculation, which predicted an activation energy of
12.9 kcal molÀ1 (see ESIw). We further analyzed model compound
9, with a coalescence temperature TC of À5 1C being observed
(Fig. 2B and Fig. S3B (ESIw)), leading to the evaluation of the free
energy of activation DG of 11.4 Æ 0.2 kcal molÀ1. Notably,
this value is again in agreement with the calculated value
(11.6 kcal molÀ1) (see ESIw). Because –S– and –CH2– moieties have
similar sizes, steric hindrance was expected to be the only effective
contributing factor in the case of 9. The difference in the DG values
obtained for model compounds 6 and 9 could then be used to
5
6
7
8
9
a
All the spectra were run under identical conditions and were
recorded at 564.6 MHz on Varian Inova-600 spectrometers. The
chemical shifts are reported in parts per million (ppm) with respect
to CF3COOH used as an external reference.
free rotation of the fluorinated phenyl ring around the C1–C2
bond of the thioacetal compounds (Fig. 2). This phenomenon,
in turn, gives rise to the chemical non-equivalence of the
F atoms and results in the observed unusual 19F NMR signals.
The FÁÁÁS interaction could be ascribed to steric hindrance
resulting either from the large size of the S atom or from an orbital
interaction between the highly polarizable S atom and the
extremely electronegative F atom, or a combination of both. To
disentangle these contributions, we investigated three other model
compounds (7–9) designed to modify the steric impediment.
estimate the strength of the FÁÁÁS interaction, i.e., 1.4 kcal molÀ1
.
This value is considerably stronger than conventional van der Waals
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 4284–4286 4285

View Article Online
Fig. 3 Schematic representation of the electron delocalization from
the fluorine lone pair (nF) (green–orange) to the anti-bonding orbital
of a S–C bond (s*S–C) (yellow–blue) in 6. Atom color code: C, gray;
H, white; F, cyan; S, yellow.
Fig. 4 Examples with the FÁ Á ÁS distance shorter than 3.27 A, the sum
of van der Waals radii of F and S. (A) is a small molecular example;
(B) is a protein/ligand complex with the 3-fluorotyrosine in the mutant
glutathione (GSH) transferase. Atom color code: C, gray; F, cyan;
S, yellow; Cl, green; N, blue; O, red.
forces (0.2–0.6 kcal molÀ1) and yet in the range of weak H-bonds.
Therefore, the FÁÁÁS interaction could be considered as a weak
interaction.
protein complexes. Fig. 4B shows that the F atom of 3-fluoro-
tyrosine in the mutant glutathione transferase lies very close to the
S atom in the ligand (3.20 A).12 Collectively, all these data provide
multiple evidence for the FÁÁÁS interaction, which might exist
ubiquitously and bring about considerable changes in structure
and properties such as those described above.
To further test the hypothesis that the FÁÁÁS interaction might
arise from an orbital interaction, we performed ab initio mole-
cular orbital (MO) calculations and natural bond orbital (NBO)
analysis6 on compounds 6 and 9. The fully optimized geometries
of 6 and 9 were obtained (Fig. S1, ESIw). Interestingly, the global
energy minimum conformation for 6 (Fig. S1A, ESIw) had one
FÁÁÁS with inter-atomic distance (3.19 A) shorter than the van der
Waals contact, the alternative FÁÁÁS distance being much larger
(3.55 A). In the case of 9, this asymmetry was not detected
(Fig. S1B, ESIw). The estimated value for the NBO delocalization
energy DEdel was around 3.5 kcal molÀ1 (Table S1, ESIw) for 6,
suggesting that a non-negligible FÁÁÁS nonbonded interaction
could arise from the orbital interaction between the divalent S
moiety (the S–C bond) and the fluorine atom (Fig. 3). Notably, in
It is worth mentioning that various weak interactions involving
fluorine have recently attracted increasing interest following the
flourishing development of fluorinated compounds in materials
science and medicinal chemistry. Among them, the seemingly
weak interactions often contribute critically to the special and
unique properties manifested by some fluorinated compounds.1a,c
The weak FÁÁÁS interaction disclosed here has been largely
ignored until now and has not received thorough investigation
yet. Future studies on this and other fluorine interactions will
undoubtedly enhance our understanding of the special properties
demonstrated by fluorinated compounds, for which certain
peculiar phenomena are often left unexplained.
a recent work, Gabbaı and Zhao obtained quite similar results by
¨
performing NBO analysis on a zwitterionic sulfonium fluoro-
borate; in this case, the same orbital interaction between S and F
was invoked to explain both the stability and the reactivity of the
considered fluoroborate molecule.7 Also, Allegra et al. estimated,
via ab initio calculations, that the presence of FÁÁÁS interactions in
substituted bithiophenes affected the conformations of these
compounds and that the intensity of such interactions is in the
range of H-bonds.8 It is worth noting that the significantly large
changes in the occupancies are only observed for the s* orbital of
the S–C bond (s*S–C) and the nF natural orbital of fluorine
(Table S1, ESIw) with marginal changes for the bonding orbital of
the S–C bond (sS–C) and the Rydberg orbital on S. These results
clearly demonstrate that the orbital interaction between S and F
may be the predominant FÁÁÁS nonbonded interaction, its main
origin being the electron delocalization from the fluorine lone pair
to the low-lying s* orbital of a C–S bond. Interestingly, another
theoretical study carried out on ortho substituted arylselenides
proposed that the n (from electron-pair donor) and s* (from
electron-pair acceptor) orbital overlap is indeed a contributing
factor toward the intramolecular nonbonding interaction.9
Considering the few reports on the FÁÁÁS weak interaction,7–10
we next undertook a survey at the Cambridge Crystallographic
Data Center (CCDC) which allowed us to find a considerable
number of high resolution crystal structures exhibiting inter- and
intramolecular FÁÁÁS distances shorter than the sum of van der
Waals radii of F and S (3.27 A) (Fig. S2, ESIw). Fig. 4A represents
an example of such a structure in which the F atom is situated close
to the S atom (3.06 A).11 This is likely due to an attractive FÁÁÁS
interaction as opposed to the repulsive steric interaction. By further
searching PDB, we also found that the FÁÁÁS interactions existed in
We thank Dr Michel Giorgi for his help during the CCDC
crystal structure survey.
Notes and references
1 (a) K. Muller, C. Faeh and F. Diederich, Science, 2007, 317, 1881;
¨
(b) K. Reichenbacher, H. I. Suss and J. Hulliger, Chem. Soc. Rev.,
2005, 34, 22; (c) S. Purser, P. R. Moore, S. Swallow and
V. Gouverneur, Chem. Soc. Rev., 2008, 37, 320; (d) S. G. Dimagno
and H. Sun, Curr. Top. Med. Chem., 2006, 6, 1473.
2 J. M. Delfino, S. L. Schreiber and F. M. Richards, J. Am. Chem.
Soc., 1993, 115, 3458.
3 (a) Q. Peng, Y. Xia, F. Qu, X. Wu, D. Campese and L. Peng,
Tetrahedron Lett., 2005, 46, 5893; (b) Y. Xia, F. Qu, A. Maggiani,
K. Sengupta, C. Liu and L. Peng, Org. Lett., 2011, 13, 4248.
4 (a) J. F. W. Keana and S. X. Cai, J. Org. Chem., 1990, 55, 3640;
(b) K. A. Chehade, K. Kiegiel, R. J. Isaacs, J. S. Pickett,
K. E. Bowers, C. A. Fierke, D. A. Andres and H. P. Spielmann,
J. Am. Chem. Soc., 2002, 124, 8206.
5 P. Lesot, O. Lafon, H. B. Kagan and C.-A. Fan, Chem. Commun.,
2006, 389.
6 A. E. Reed, L. A. Curtiss and F. Weinhold, Chem. Rev., 1988,
88, 899.
7 H. Zhao and F. P. Gabbaı, Org. Lett., 2011, 13, 1444.
¨
8 G. Raos, A. Famulari, S. V. Meille, M. C. Gallazzi and G. Allegra,
J. Phys. Chem. A, 2004, 108, 691.
9 D. Roy and R. B. Sunoj, J. Phys. Chem. A, 2006, 110, 5942.
10 (a) C. R. Wade, H. Zhao and F. P. Gabbaı, Chem. Commun., 2010,
¨
46, 6380; (b) Y. Wang, S. R. Parkin, J. Gierschner and
M. D. Watson, Org. Lett., 2008, 10, 3307.
11 F. Nicolo, G. Bruno, R. Scopelliti, S. Grasso, A. Rao and M. Zappala,
Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2001, 57, 572.
12 G. Xiao, J. F. Parsons, R. N. Armstrong and G. L. Gilliland,
J. Am. Chem. Soc., 1997, 119, 9325.
c
4286 Chem. Commun., 2012, 48, 4284–4286
This journal is The Royal Society of Chemistry 2012