Discrimination of chain positions in mixed short/long-chain glycerophosphocholines by NMR chemical shift variations

Article abstract of DOI:10.1007/s11746-008-1280-4

The synthesis of a series of (1,2-) mixed short/long-chain glycerophosphocholines has been performed. Starting from glycerophosphorylcholine (GPC), and using regioselective acylation in the presence of dibutyltin oxide, a set of high-purity isomeric mixed-chain phospholipids was obtained. This has allowed the development of a simple NMR method for the structural determination of the isomeric 1(2)-short-2(1)-long- diacylglycerophosphocholines. The method is based on the observation that selected protons in the two series of isomeric phospholipids undergo systematic chemical shift variations Δδ that can be ascribed to the acyl substituents on the glycerol backbone. The observed patterns can be exploited as a simple method for the discrimination of regioisomeric unsymmetrical 1,2-diacylglycerophosphocholines.

Full text of DOI:10.1007/s11746-008-1280-4

J Am Oil Chem Soc (2008) 85:1005–1011
DOI 10.1007/s11746-008-1280-4
ORIGINAL PAPER
Discrimination of Chain Positions in Mixed Short/Long-Chain
Glycerophosphocholines by NMR Chemical Shift Variations
Paola D’Arrigo Æ Andrea Mele Æ Cristina Rossi Æ
Davide Tessaro Æ Stefano Servi
Received: 27 March 2008 / Revised: 23 June 2008 / Accepted: 19 July 2008 / Published online: 3 September 2008
Ó AOCS 2008
Abstract The synthesis of a series of (1,2-) mixed short/
long-chain glycerophosphocholines has been performed.
Starting from glycerophosphorylcholine (GPC), and using
regioselective acylation in the presence of dibutyltin oxide,
a set of high-purity isomeric mixed-chain phospholipids
was obtained. This has allowed the development of a
simple NMR method for the structural determination of
the isomeric 1(2)-short-2(1)-long-diacylglycerophospho-
cholines. The method is based on the observation that
selected protons in the two series of isomeric phospholipids
undergo systematic chemical shift variations Dd that can be
ascribed to the acyl substituents on the glycerol backbone.
The observed patterns can be exploited as a simple method
for the discrimination of regioisomeric unsymmetrical
Although lysophospholipids are well-known PL metabo-
lites, their role as intercellular signaling molecules is being
studied in more and more depth [3, 4]. Lysophospholipids
and their receptors have been found in a great range of
tissues and cell types, and are involved in many processes
such as reproduction, angiogenesis, and nervous system
regulation [5, 6].
2-Acyl-1-lyso-sn-glycerophosphocholines are very inter-
esting biologically active compounds because they are
potential in vivo delivery systems for chains situated at
position sn-2. It has recently been shown [7] that 2-doc-
osahexaneoyl-1-lysophosphatidylcholine is preferentially
transferred through the in vitro blood–brain barrier over
unesterified docosahexanenoic acid. Nevertheless, these
compounds are not stable in vivo due to spontaneous acyl
chain migration from position sn-2 to the more stable sn-1
[8].
1
,2-diacylglycerophosphocholines.
Keywords Regioselective acylation ꢀ
Diacylglycerophosphocholine ꢀ Isomeric phospholipids ꢀ
Mixed chains ꢀ NMR
It has been stated [9] that the presence of a small acyl
chain in position sn-1 causes the PL to display the same
biochemical behavior as 1-lysoPL, because the sn-1 posi-
tion is occupied and not free for spontaneous acyl
migration from the less stable sn-2 position. We have also
recently reported [10] a practical synthesis of mixed short/
long 1,2-diacyl-chain phosphatidylcholines starting from
glycerophosphorylcholine (GPC). Figure 1 shows the
general formulae and atom numbering of the newly syn-
thesized compounds.
Introduction
Recently, research has increasingly been focused on bio-
logically active phospholipids (PLs), and in particular on
lysophospholipids (LPLs), due to their active role in many
biological processes [1, 2]. LPLs are widely distributed in
nature in both animals and plants, even though they rep-
resent only a small fraction of the cellular lipid component.
During the preparation of mixed-chain PLs from GPC,
we had the opportunity to compare different sets of 1,2-
phosphatidylcholine (PC) isomers, as formally obtained by
swapping the same pairs of substituents (e.g., C2 and C16
fatty acid residues) from position 1 to position 2 of the
glycerol backbone and vice versa.
P. D’Arrigo (&) ꢀ A. Mele ꢀ C. Rossi ꢀ D. Tessaro ꢀ S. Servi
Dipartimento di Chimica, Materiali e Ingegneria Chimica G.
Natta, Politecnico di Milano, Via Mancinelli 7,
The determination of fatty acid composition in phos-
pholipids is usually obtained either by selective enzymatic
20131 Milan, Italy
e-mail: paola.d’arrigo@polimi.it
1
23

1
006
J Am Oil Chem Soc (2008) 85:1005–1011
as those labeled as H -H and H and H in Fig. 2. The
A
E
a
b
observed chemical shift turned out to be dependent upon
the chain length and the position of the acyl substituent on
the glycerol backbone. Based on these observations, we
propose a simple nondestructive method for the easy dis-
crimination of regioisomers based on chemical shift data
for selected protons without any chemical manipulations.
Experimental Procedures
Materials and Methods
All chemicals were purchased from Sigma–Aldrich (St.
Louis, MO, USA). All solvents were of analytical grade.
?
The ion exchange resin is an H -form cation exchange
resin (Fluka 44443, Buchs, Switzerland), 20–50 mesh.
1
H NMR spectra were recorded on Bruker (Rheinstetten,
Germany) ARX 400 and AV 500 instruments operating at
1
H resonance frequencies of 400 and 500 MHz, respec-
Fig. 1 Molecular formulae and atom numbering of compounds 1–12
tively. Chemical shifts (d, ppm) are reported relative to
tetramethylsilane (TMS) used as internal standard. All
spectra were recorded in CDCl /CD OD (see ‘‘Synthesis’’)
3
3
at 305 K. The assignment of proton spectra was supported
by standard 2D correlation experiments (COSY and
TOCSY) obtained by using library pulse sequences.
Nuclear Overhauser enhancement correlation experiments
(NOESY) were carried out by using standard pulses in the
literature. The spectra were acquired by using magnetic
field gradients for coherence pathway selection. The data
were collected with three different mixing times, 300, 600,
and 900 ms.
Mass spectra were recorded on an ESI/MS (Bruker
Esquire 3000) by direct infusion of methanol solutions of
compounds 1–12.
Silica gel 60 F254 plates (Merck, Darmstadt, Germany)
were used for analytical TLC; mixtures of CHCl /CH OH/
1
Fig. 2 Expansions of the H NMR spectra of compounds 11 (top
trace) and 12 (bottom trace)
3
3
NH in different ratios were used as eluant. Detection was
3
achieved with UV light followed by I staining.
2
hydrolysis in positions sn-1 and sn-2 with some lipases [11]
and phospholipase A [12, 13], respectively, and the sub-
HPLC analysis were performed on an Agilent (Palo
Alto, CA, USA) 1100 series apparatus fitted with a
LiChrospher (Merck) 100 diol 5 lm column, length/inter-
nal diameter = 125/4, and an evaporative light scattering
detector (ELSD) (Sedex model 75, Sedere, Lawrenceville,
NJ, USA, 41 °C). The operating temperature and pressure
were 55 °C and 3.5 bar. A binary solvent system of solvent
A: hexane/isopropanol/acetic acid/triethylamine (815/170/
15/0.8, v/v/v/v) and solvent B: isopropanol/water/acetic
acid/triethylamine (837/140/15/0.8, v/v/v/v) was used in a
gradient mode starting from 60% A, 40% B and ramping to
100% B in 16 min. The 20 ll samples of a 1 mg/ml
CH Cl /CH OH solution were injected into the column at a
2
sequent analysis of released fatty acid residue [14], or by
using liquid chromatography/electrospray ionization/mass
spectrometry [15]. In all cases the analyzed sample cannot
recovered.
In the present work, the long- and short-chain compo-
nents of the PLs were palmitic acid and short-chain
saturated linear acid residues ranging from C2 to C10,
respectively. The analysis and discrimination of regioi-
somers is hampered by the fact that the homologous pairs
exhibit similar retentions in both HPLC and TLC analysis.
1
Nevertheless, the analysis of H NMR spectra showed the
2
2
3
existence of diagnostic signals for each isomeric pair, such
mobile phase flow rate of 1.5 ml/min.
1
23

J Am Oil Chem Soc (2008) 85:1005–1011
Synthesis
1007
1
The H NMR spectrum has been already reported in the
?
literature [11]. ESI/MS: [M ? H] = 496.3; [M ? Na] =
?
The synthesis of the isomers reported in Fig. 1 was carried
out with the same procedure. We report here as an example
the detailed synthesis of the isomeric compounds 1 and 2
and analytical data for all compounds.
518.2. t (HPLC) = 6.24 min.
R
1-O-Hexadecanoyl-2-O-acetyl-sn-glycero-3-
phosphocholine (2)
To a solution of 1-O-hexadecanoyl-sn-glycero-3-phospho-
choline (200 mg, 0.404 mmol) in dry dichloromethane
1-O-Acetyl -2-O- hexadecanoyl-sn-glycero-3-
phosphocholine (1)
(
(
8 ml), DMAP (148 mg, 1.21 mmol) and acetic anhydride
0.112 ml, 1.21 mmol) were added. The reaction mixture
GPC (250 mg, 0.97 mmol) and dibutyltin oxide
was stirred for 20 h at r.t., then vacuum-distilled. The crude
product was then dissolved in dichloromethane and stirred
at room temperature for 1 h with Dowex 50 9 8 cation
(
(
DBTO;275 mg, 1.1 mmol) were suspended in 2-propanol
10 ml) and refluxed for 1 h. The mixture was cooled to 25
°
C. Then dimethylamino pyridine (DMAP; 147 mg, 1.2
?
exchange resin (H -form). The suspension was then
mmol) and acetyl chloride (83.2 ll, 1.2 mmol) were added.
After 48 h, the suspension was filtered and the solution was
concentrated under vacuo without heating. The residue was
suspended in dry dichloromethane (10 ml) and treated with
DMAP (0.335 g, 2.74 mmol) and palmitic anhydride (1.36
g, 2.74 mmol). The reaction mixture was stirred for 24 h at
r.t., then concentrated under vacuo. The crude product was
then dissolved in heptane (10 ml) and water (10 ml). After
centrifugation the organic phases were evaporated to give a
residue which was dissolved in dichloromethane (2 ml) and
precipitated in acetone (30 ml) at -20 °C. Solid 1 (160 mg,
filtered and concentrated to give a residue which was dis-
solved in dichloromethane (2 ml) and precipitated in
acetone (30 ml) at -20 °C. Solid 2 (141 mg, 65% yield)
was recovered by filtration and dried under vacuo.
1
H NMR (CDCl /CD OD): d 5.20 (1H, m, H ), 4.36
3
3
C
(3H, m, H ? H ), 4.15 (1H, m, H ), 3.99 (2H, m, HD/E),
B a A
3.80 (2H, m, H ), 3.34 (9H, s, N(CH ) ), 2.27 (2H, m,
b 3 3
COCH ), 2.05 (3H, s, COCH ), 1.58 (2H, m, COCH CH ),
2
3
2
2
1
.26 (24H, m, COCH CH (CH ) CH ), 0.88 (3H, m,
2 2 2 12 3
?
CO(CH ) CH ). ESI/MS: [M ? H] = 538.4; [M ? Na]
?
2 14
3
=
560.4. t (HPLC) = 5.305 min.
R
3
0 %) was recovered by filtration and dried under vacuo.
1
H NMR (CDCl /CD OD): d 5.15 (1H, m, H ), 4.33
3
3
C
1-O-Propanoyl -2-O- hexadecanoyl-sn-glycero-3-
phosphocholine (3)
(
1H, m, H ), 4.18 (2H, m, H ), 4.07 (1H, m, H ), 3.93 (2H,
B a A
m, HD/E), 3.52 (2H, m, H ), 3.11 (9H, s, N(CH ) ), 2.22
3 3
b
(
2H, m, COCH ), 2.03 (3H, s, COCH ), 1.54 (2H, m,
2 3
1
H NMR (CDCl /CD OD): d 5.16 (1H, m, H ), 4.33 (1H,
3
3
C
COCH CH ), 1.18 (24H, m, COCH CH (CH ) CH ),
3
2
2
2
2
2 12
?
m, H
D/E), 3.60 (2H, m, H
m, COCH ), 1.61 (2H, m, COCH CH ), 1.25 (24H, m,
B
), 4.22 (2H, m, H
a
), 4.09 (1H, m, H
A
), 3.95 (2H, m,
0
.82 (3H, m, CO(CH ) CH ). ESI/MS: [M ? H] = 538.2;
2 14 3
?
H
b
), 3.25 (9H, s, N(CH
) ), 2.32 (4H,
3 3
[
M ? Na] = 560.2. t (HPLC) = 5.178 min.
R
2
2
2
COCH CH (CH ) CH ), 0.98 (6H, m, CO(CH ) CH ).
2
2
2 12
3
2 n
3
1
-O-Hexadecanoyl-sn-glycero-3-phosphocholine
1-O-PalmPC)
?
?
ESI/MS: [M ? H] = 552.4; [M ? Na] = 574.4. tR
HPLC) = 4.682 min.
(
(
GPC (1 g, 3.9 mmol) and DBTO (1.1 g, 3.9 mmol) were
suspended in 2-propanol (40 ml) and refluxed for 1 h. The
mixture was cooled to 25 °C and treated with triethyl-
amine (TEA; 0.65 ml, 4.7 mmol) and palmitoyl chloride
1-O-Hexadecanoyl-2-O-propanoyl-sn-glycero-3-
phosphocholine (4)
1
H NMR (CDCl /CD OD): d 5.21 (1H, m, H ), 4.36 (3H, m,
3
3
C
(
1.28 g, 4.7 mmol). The formation of the lysoPC was
H ? H ), 4.14 (1H, m, H ), 3.99 (2H, m, HD/E), 3.85 (2H, m,
A
B
a
followed by HPLC. After 15 min the reaction mixture
contained 97% of 1-palmitoyl-2-lyso-phosphatidylcholine
and 3% of GPC as the only phospholipids. The solution
was then treated with water (40 ml) and extracted with
heptane (40 ml). The water–alcohol solution was extrac-
ted three times with heptane (3 9 40 ml) and evaporated
in vacuo; the crude residue was dissolved in ethanol (10
ml) and precipitated with acetone (40 ml) at -10 °C to
give 1.5 g of 1-O-Palm-PC as a white powder (80%). The
analytical samples were obtained after crystallization from
ethanol (yield 40%).
H ), 3.38 (9H, s, N(CH ) ), 2.28 (4H, m, COCH ), 1.38 (2H,
b
3 3
2
m, COCH CH ), 1.25 (24H, m, COCH CH (CH ) CH ),
2 12
2
2
2
2
3
1
.15 (3H, m, COCH CH ), 0.88 (3H, m, CO(CH ) CH ).
2 3 2 14 3
?
ESI/MS: [M ? H] = 552.2; [M ? Na] = 574.2. t (HPLC) =
?
R
4.582 min.
1-O-Butanoyl -2-O-hexadecanoyl-sn-glycero-3-
phosphocholine (5)
1
H NMR (CDCl /CD OD): d 5.20 (1H, m, H ), 4.37 (1H,
3
3
C
m, H ), 4.29 (2H, m, H ), 4.11 (1H, m, H ), 3.97 (2H, m,
B
a
A
1
23

1
008
J Am Oil Chem Soc (2008) 85:1005–1011
HD/E), 3.70 (2H, m, H ), 3.22 (9H, s, N(CH ) ), 2.31 (4H,
b 3 3
H ), 3.28 (9H, s, N(CH ) ), 2.23 (4H, m, COCH ), 1.53 (4H,
b
3 3
2
m, COCH ), 1.61 (4H, m, COCH CH ), 1.23 (24H, m,
2
m, COCH CH ), 1.20 (34H, m, COCH CH (CH ) CH ),
2 2 2 2 2 n 3
2
2
?
COCH CH (CH ) CH ), 0.87 (6H, m, CO(CH ) CH ).
3
0.81 (6H, m, CO(CH ) CH ). ESI/MS: [M ? H] = 636.5;
2 n 3
2
2
2 12
3
2 n
?
ESI/MS: [M ? Na] = 588.3. t (HPLC) = 4.354 min.
?
[M ? Na] = 658.5. t (HPLC) = 3.609 min.
R
R
1-O-Hexadecanoyl-2-O-butanoyl-sn-glycero-3-
phosphocholine (6)
1-O-Decanoyl-2-O-hexadecanoyl-sn-glycero-3-
phosphocholine (11)
1
1
H NMR (CDCl /CD OD): d 5.19 (1H, m, H ), 4.34 (1H, m,
H NMR (CDCl /CD OD): d 5.13 (1H, m, H ), 4.30 (1H,
3
3
C
3
3
C
H ), 4.28 (2H, m, H ), 4.11 (1H, m, H ), 3.98 (2H, m, HD/E),
A
m, H ), 4.18 (2H, m, H ), 4.05 (1H, m, H ), 3.91 (2H, m,
B a A
B
a
3
.68 (2H, m, H ), 3.25 (9H, s, N(CH ) ), 2.26 (4H, m,
b 3 3
HD/E), 3.55 (2H, m, H ), 3.15 (9H, s, N(CH ) ), 2.23 (4H,
b 3 3
COCH ), 1.63 (4H, m, COCH CH ), 1.22 (24H, m,
2
m, COCH ), 1.55 (4H, m, COCH CH ), 1.17 (36H, m,
2 2 2
2
2
COCH CH (CH ) CH ), 0.93 (3H, m, CO(CH ) CH ). ESI/
3
COCH CH (CH ) CH ), 0.80 (6H, m, CO(CH ) CH ).
2 2 2 n 3 2 n 3
2
2
2 12
3
2 n
?
MS: [M ? H] = 566.3; [M ? Na] = 588.3. t (HPLC) =
?
?
ESI/MS: [M ? H] = 650; [M ? Na] = 672. t (HPLC) =
?
R
R
4
.434 min.
3.65 min.
1-O-Hexanoyl-2-O-hexadecanoyl-sn-glycero-3-
phosphocholine (7)
1-O-Hexadecanoyl-2-O-decanoyl-sn-glycero-3-
phosphocholine (12)
1
1
H NMR (CDCl /CD OD): d 5.12 (1H, m, H ), 4.30 (1H,
H NMR (CDCl /CD OD): d 5.15 (1H, m, H ), 4.30 (3H, m,
3
3
C
3
3
C
m, H ), 4.20 (2H, m, H ), 4.04 (1H, m, H ), 3.90 (2H, m,
B
H ? H ), 4.07 (1H, m, H ), 3.98 (2H, m, HD/E), 3.74 (2H, m,
B a A
a
A
HD/E), 3.58 (2H, m, H ), 3.18 (9H, s, N(CH ) ), 2.23 (4H,
3 3
H ), 3.24 (9H, s, N(CH ) ), 2.23 (4H, m, COCH ), 1.52 (4H,
b 3 3 2
b
m, COCH ), 1.64 (4H, m, COCH CH ), 1.16 (28H, m,
2
m, COCH CH ), 1.18 (36H, m, COCH CH (CH ) CH ),
2 2 2 2 2 n 3
2
2
?
COCH CH (CH ) CH ), 0.81 (6H, m, CO(CH ) CH ).
2 n
0.80 (6H, m, CO(CH ) CH ). ESI/MS: [M ? H] = 650; [M
2 n 3
2
2
2 n
3
3
?
?
?
? Na] = 672.4. t (HPLC) = 3.71 min.
ESI/MS: [M ? H] = 594.4; [M ? Na] = 616.4. tR
HPLC) = 3.941 min.
R
(
1-O-Hexadecanoyl-2-O-hexanoyl-sn-glycero-3-
phosphocholine (8)
Results and Discussion
We have recently described [10] a new procedure for the
synthesis of mixed short/long-chain glycerophospho-
cholines which exploits the tin-mediated synthesis of
lysophospholipids [16] and subsequent selective acylation
in the sn-2 position of the glycerol backbone.
1
H NMR (CDCl /CD OD): d 5.23 (1H, m, H ), 4.38 (3H, m,
3
3
C
H ? H ), 4.14 (1H, m, H ), 4.02 (2H, m, HD/E), 3.94 (2H, m,
B
a
A
H ), 3.34 (9H, s, N(CH ) ), 2.26 (4H, m, COCH ), 1.58 (4H,
3 3
b
2
m, COCH CH ), 1.26 (28H, m, COCH CH (CH ) CH ),
2
2
2
2
2 n
3
?
.88 (6H, m, CO(CH ) CH ). ESI/MS: [M ? H] = 594.3;
0
In this procedure, glycerophosphocholine (GPC) is
transformed into the cyclic stannylene derivative by treat-
ment in 2-propanol with dibutyltin oxide (DBTO) at reflux.
The cyclic tin ketal is then acylated with the desired acid
chloride and dimethylaminopyridine (DMAP) as a base.
The solvent was then removed at room temperature and
replaced with dry dichloromethane. The second acylation is
carried out with the acid anhydride in the presence of fresh
DMAP. After different purification steps, yields from GPC
were in the range of 20 to 69% for compounds 1–12. This
tin-catalyzed synthesis of phospholipids always leads to a
single regioisomer. The purities of the compounds were
investigated by HPLC and their structures were determined
2
n
3
?
[
M ? Na] = 616.3. t (HPLC) = 4.195 min.
R
1-O-Nonanoyl-2-O-hexadecanoyl-sn-glycero-3-
phosphocholine (9)
1
H NMR (CDCl /CD OD): d 5.14 (1H, m, H ), 4.32 (1H,
3
3
C
m, H ), 4.19 (2H, m, H ), 4.06 (1H, m, H ), 3.92 (2H, m,
A
B
a
HD/E), 3.53 (2H, m, H ), 3.25 (9H, s, N(CH ) ), 2.28 (4H,
b
3 3
m, COCH ), 1.60 (4H, m, COCH CH ), 1.23 (34H, m,
2
2
2
COCH CH (CH ) CH ), 0.85 (6H, m, CO(CH ) CH ).
2
?
ESI/MS: [M ? H] = 636.5; [M ? Na] = 658.5.
2
2 n
3
2 n
3
?
tR (HPLC) = 3.51 min.
1
by H NMR.
1
-O-Hexadecanoyl-2-O-nonanoyl-sn-glycero-3-
phosphocholine (10)
For the sake of clarity, compounds 1–12 can be divided
into two families of isomers: odd-numbered compounds
correspond to PCs carrying a variable length acyl sub-
stituent at position 1 of GPC and a palmitoyl moiety at
position 2, whilst even-numbered compounds correspond
1
H NMR (CDCl /CD OD): d 5.16 (1H, m, H ), 4.30 (3H, m,
3
3
C
H ? H ), 4.07 (1H, m, H ), 3.95 (2H, m, HD/E), 3.76 (2H, m,
a
B
A
1
23

J Am Oil Chem Soc (2008) 85:1005–1011
1009
to their regioisomeric counterparts, namely those PCs
bearing the palmitoyl chain at position 1 and the same acyl
group as before at position 2. Accordingly, short indicative
names for the isomeric pairs can be used, such as C2-
PalmPC and Palm-C2PC for compounds 1 and 2,
respectively.
The NMR spectra of compounds 1–12 showed the
presence of isolated signals in the chemical shift range
5
.40–3.60 ppm, along with the envelope due to nonassign-
able protons of the hydrocarbon chains, the terminal methyl
groups and the singlet of the three equivalent CH groups of
3
1
the choline moiety. Expansions of the H-NMR spectra for
the homologous pair 11–12 are reported in Fig. 2.
The protons belonging to the glycerol frame (see Fig. 1)
are expected to give rise to an ABCMX spin system. As a
matter of fact, the protons HD/E attached to C3 of GPC are
isochronous and thus the observed spin system is ABCM2.
Any attempt to remove their chemical shift degeneracy by
changing solvent and/or temperature failed. The NMR
signal of the latter protons then appears as a deceptively
Fig. 3 Reported differences in chemical shift (Dd, ppm) for selected
protons in compounds 1–12 vs the number Cn of carbon atoms in the
variable length acyl chain. Dd for a given proton is defined as d(H in
Palm-CnPC)-d(H in Cn-PalmPC) for all of the examined Cn
simple triplet. In a similar way, the CH groups of the
2
latter in the GPC backbone; (2) the Dd values are always
positive, with the exception of the pair 5–6 (C4-PalmPC
choline fragment gave rise to a triplet (CH a) and to a
2
broad singlet (CH b).
2
and Palm-C4PC); and (3) the values for the choline CH2
The chemical shifts of H -H , H and H were collected
A
E
a
b
protons are one order of magnitude larger than all of the
and tabulated for all of the examined compounds. The
results are summarized in Table 1.
others. A graphical summary is reported in Fig. 3.
The chemical shift data clearly indicate that both the
glycerol protons and the choline methylene groups are
sensitive probes of regioisomeric substitution and useful
probes for recognizing the acyl substitution pattern at
the C1–C2 frame of GPC. The large Dd observed for
the choline methylene groups are far beyond the exper-
imental uncertainty and suggest that this quantity could
be used as an unbiased descriptor for discriminating
the regioisomeric substitution of mixed short/long-chain
glycerophosphocholines.
Some systematic trends could be better highlighted by
calculating the difference in the chemical shift for the same
proton between the homologous isomeric pairs. The results
are reported in the Dd columns of Table 1.
For example, the column DdH reports the difference in
C
the chemical shift of HC between the Palm-CnPC com-
pounds and the corresponding Cn-PalmPC (Cn = C2, C3,
C4, C6, C9, C10). After studying the Dd values, it is clear
that: (1) the chemical shift variations can be related to the
length of the acyl substituent and to the position of the
Table 1 Chemical shifts (d, ppm) and the differences in chemical shift (Dd, ppm) for selected protons from compounds 1–12
0
00
Compd.
R
R
Short name
H
C
DdH
C
H
B
DdH
B
H
A
DdH
A
H
D/E
DdHD/E
H
a
DdH
a
H
b
DdH
b
1
2
3
4
5
6
7
8
9
C2 C16 C2-PalmPC 5.150
–
4.330
–
4.069
–
3.925
–
4.177
–
3.524
–
C16 C2 Palm-C2PC 5.205 0.055
C3 C16 C3-PalmPC 5.160
C16 C3 Palm-C3PC 5.214 0.054
C4 C16 C4-PalmPC 5.200
C16 C4 Palm-C4PC 5.193 -0.007 4.343 -0.027 4.106 –0.008 3.977 0.003
C6 C16 C6-PalmPC 5.120 4.303 4.040 3.901
C16 C6 Palm-C6PC 5.227 0.107 4.383 0.08 4.136 0.096 4.022 0.121
C9 C16 C9-PalmPC 5.136 4.316 4.061 3.920
C16 C9 Palm-C9PC 5.159 0.023 4.298 -0.018 4.073 0.012 3.950 0.03
C10 C16 C10-PalmPC 5.126 4.303 4.053 3.906
C16 C10 Palm-C10PC 5.152 0.026 4.296 -0.007 4.067 0.014 3.981 0.075
4.360 0.03
4.330
4.364 0.034
4.370
4.150 0.081 3.999 0.074
4.090 3.950
4.140 0.05 3.999 0.049
4.114 3.974
4.360 0.183
4.220
4.364 0.144
4.294
4.275 -0.019 3.683 -0.015
4.201 3.582
4.383 0.182 3.936 0.354
4.185 3.532
4.300 0.115 3.755 0.223
4.175 3.549
4.296 0.121 3.736 0.187
3.802 0.278
3.600
3.845 0.245
3.698
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
1
1
1
0
1
2
–
–
–
–
–
–
1
23

1
010
J Am Oil Chem Soc (2008) 85:1005–1011
3
Table 2 Values (Hz) of the vicinal ( J) coupling constants for the
examined compounds
In order to provide a possible rationale for the empirical
criterium reported above, a conformational investigation
of compounds 1–12 was undertaken. The conformational
preferences of the examined compounds can be described
in terms of two dihedral angles, h and u, corresponding to
the torsions around the C1–C2 and C2–C3 chemical bonds
of GPC, respectively. In turn, h and u drive the values of
vicinal proton spin–spin coupling constants via the
Karplus–Altona equation [17]. It should be noted that, in
the present case, the tremendous overlap of the signals
0
00
Compound
R
R
Short name
J
BC
J
AC
J
R = JAC/JBC
1
2
3
4
5
6
7
8
9
C2
C16 C2
C3 C16 C3-PalmPC
C16 C3 Palm-C3PC
C4 C16 C4-PalmPC
C16 C4 Palm-C4PC
C6 C16 C6-PalmPC
C16 C6 Palm-C6PC
C9 C16 C9-PalmPC
C16 C9 Palm-C9PC
C16 C2-PalmPC
3.3 6.9 2.1
3.2 6.8 2.1
3.3 7.0 2.1
3.0 6.8 2.3
3.3 7.1 2.2
3.5 6.7 1.9
3.0 7.1 2.4
3.4 6.9 2.0
3.3 7.0 2.1
3.3 6.7 2.0
Palm-C2PC
0
00
from the acyl chains R and R (see Fig. 1) severely
hampers the use of inter-chain NOE for conformational
analysis. Therefore, the conformational preferences of the
title compounds can only be spotted via the analysis of J
couplings. In detail, the experimental values of JAC and JBC
are related to h, while those of JCD and JCE are related to u.
In the case of different conformations that rapidly inter-
convert on the time-scale of NMR spectroscopy, the
observed vicinal coupling constants are the average of the
coupling constants of pure conformations weighted by the
fractional population of the conformers. In our case, the
values of JCD and JCE could not be experimentally deter-
mined due to the degeneracy of the D and E protons, as
described above. Thus, our data allow us to discuss the
conformational preferences of 1–12 concerning the torsion
around the C1–C2 bond of GPC only. A sketch of the
rotamers that are appreciably populated is reported in
Fig. 4.
1
0
11
C10 C16 C10-PalmPC 3.3 7.0 2.1
C16 C10 Palm-C10PC 3.5 6.8 1.9
12
a
a
a
C3PC
C4PC
C6PC
a
C3
C4
C6
C3
C4
C6
3.5 6.6 1.9
3.4 6.7 2.0
3.2 7.0 2.2
From [18]
For the conformational analysis, we took advantage
of the work published by Martin and Pitzer [18] on
symmetrically substituted 1,2-dipropanoyl-, 1,2-dibuta-
noyl- and 1,2-dihexanoylphosphatidylcholines (C3PC,
C4PC, and C6PC, respectively). Those substrates gave
high-resolution NMR spectra in D O solution without any
2
J
Fig. 5 Plot of R = JAC/JBC versus the number Cn of carbon atoms in
the variable-length acyl chain
chemical shift degeneracy, and thus allowed the authors to
measure all of the vicinal coupling constants related to both
h and u. The authors found that the conformation about the
C1–C2 fragment was mainly dictated by conformers A and
B, with the third conformation C providing only a minor
contribution to the overall conformer distribution. The C2–
C3 fragment of the GPC moiety was more flexible, with the
three conformations D, E and F being nearly equally
populated. The observed vicinal coupling constants of 1–12
are summarized in Table 2.
The literature values for the references C3PC, C4PC and
C6PC are also reported for the sake of comparison. The
ratio R = JAC/JBC can be used as suitable empirical
J
descriptor to compare the conformations about C1–C2 in
corresponding pairs of regioisomers. A plot of R values vs
J
the number Cn of carbon atoms in the acyl chains is shown
in Fig. 5.
Starting from the crossover point corresponding to n = 4,
the data show that the RJ values of the odd-numbered
terms (Cn-PalmPC) are systematically greater than the RJ
values of the even-numbered isomers (Palm-CnPC). This
fact, in turn, corresponds to a higher contribution of the
A conformation with respect to the C conformation in the
Cn-PalmPC series compared to the Palm-CnPC.
Fig. 4 Sketch of populated rotamers resulting from torsion around
the C1–C2 bond of GPC. Rotamer nomenclature are as given in [18]
1
23

J Am Oil Chem Soc (2008) 85:1005–1011
1011
Therefore, significant chemical shift differences Dd
observed for the choline CH groups in unsymmetrical 1,2-
2-docosahexaneoyl-1-lysophosphatidylcholine through an in vitro
blood–brain barrier over unesterified docosahexaenoic acid.
J Neurochem 72:338–345
. Rosseto R, Hajdu J (2005) A rapid and efficient method for
migration-free acylation of lysophospholipids: synthesis of
phosphatidylcholines with sn-2-chain terminal reporter groups.
Tetrahedron Lett 46:2941–2944
. Bayon Y, Croset M, Lagarde M, Lecerf J, Thies F, Frank T,
Tayot JL, Chirouze V (1997) Docosahexaenoic acid-containing
(lyso)phosphatidylcholines for medicaments for treating cardio-
vascular disease and cerebral essential fatty acid deficiency. US
Patent 5,654,290
0. D’Arrigo P, Fasoli E, Pedrocchi-Fantoni G, Rossi C, Saraceno C,
Tessaro D, Servi S (2007) A practical selective synthesis of
mixed short/long chains glycerophosphocholines. Chem Phys
Lipids 147:113–118
2
diacyl glycerophosphocholines can be related to both the
substitution patterns at the sn-1 and sn-2 positions of the GPC
and the length of the hydrocarbon chain of the acyl moiety.
The observed patterns can be exploited as a simple, quick and
nondestructive method for discriminating regioisomeric
unsymmetrical 1,2-diacyl glycerophosphocholines. How-
ever, these chemical shift variations need to be theoretically
investigated further in order to find a general model that can
explain their behavior. Nevertheless, significant correlations
between the observed Dd values and the vicinal coupling
constants across the GPC backbone suggest that the observed
trends can be ascribed to different conformational prefer-
ences of the target molecules.
8
9
1
1
1
1. D’Arrigo P, Servi S (1997) Using phospholipases for phospho-
lipids modification. Trends Biotechnol 15:90–96
2. Mustratnta A, Forssell P, Poutanen K (1995) Comparison of
lipases and phospholipases in the hydrolysis of phospholipids.
Process Biochem 30:393–401
1
3. Six DA, Edward EA (2000) The expanding superfamily of
2
phospholipase A enzymes: classification and characterization.
Biochim Biophys Acta 1488:1–19
4. Beermann C, Mobius M, Winterling N, Schmitt JJ, Boehm G
References
1
1
2
. Birgbauer E, Chun J (2006) New developments in the biological
functions of lysophospholipids. Cell Mol Life Sci 63:2695–2701
. Goetzl EJ, An S (1998) Diversity of cellular receptors and
functions for the lysophospholipid growth factors lysophospha-
tidic acid and sphingosine 1-phosphate. FASEB J 12:1589–159
. Barenholz Y, Cevc G (2000) Structure and properties of mem-
branes. In: Baskin A, Norde W (eds) Physical chemistry of
biological interfaces. Marcel Dekker, New York, pp 171–242
. Fuller N, Rand R (2001) The influence of lysolipids on the
spontaneous curvature and bending elasticity of phospholipid
membranes. Biophys J 81(1):243–254
(
2005) sn-Position determination of phospholipid-linked fatty
acids derived from erythrocytes by liquid chromatography elec-
trospray ionization mass spectrometry. Lipids 40:211–218
5. Fang J, Barcelona MJ (1998) Structural determination and
quantitative analysis of bacterial phospholipids using liquid
chromatography/electrospray ionization/mass spectrometry.
J Microbiol Methods 33:23–35
6. Fasoli E, Arnone A, Caligiuri A, D’Arrigo P, de Ferra L, Servi S
(2006) Tin-mediated synthesis of lyso-phospholipids. Org Biomol
Chem 4:2974–2978
7. Haasnoot CAG, de Leeuw FAAM, Altona C (1980) The relation
between proton–proton NMR coupling constants and substituent
electronegativities. I. An empirical generalization of the Karplus
equation. Tetrahedron 36:2783–2792
8. Martin SF, Pitzer GE (2000) Solution conformations of short-
chain phosphatidylcholine substrates of the phosphatidylcholine-
preferring PLC of Bacillus cereus. Biochim Biophys Acta
1
3
4
5
6
7
1
1
. Muresan AS, Diamant H, Lee KY (2001) Effect of temperature
and composition on the formation of nanoscale compartments in
phospholipid membrane. J Am Chem Soc 123:6951–6952
. Ishii I, Fukushima N, Ye X, Chun J (2004) Lysophospholipids
receptors: signalling and biology. Annu Rev Biochem 73:321–
1
3
54
. Bernoud N, Fenard L, Moli e` re P, Dehouck MP, Lagarde M,
Cecchelli R, Lecerf (1999) Preferential transfer of
1
464:104–112
J
1
23