A reinvestigation of the synthesis of arsonolipids (2,3-diacyloxypropylarsonic acids)

Article abstract of DOI:10.1016/j.chemphyslip.2008.02.001

A reinvestigation of the reactions leading to arsonolipids (2,3-diacyloxypropylarsonic acids) has been carried out in order to understand why the yields of their preparation were only moderate, although they are better than those reported for 2,3-diacylox

Full text of DOI:10.1016/j.chemphyslip.2008.02.001

Available online at www.sciencedirect.com
Chemistry and Physics of Lipids 152 (2008) 113–121
A reinvestigation of the synthesis of arsonolipids
(2,3-diacyloxypropylarsonic acids)
Gerasimos M. Tsivgoulis∗, Panayiotis V. Ioannou∗∗
Department of Chemistry, University of Patras, Patras, Greece
Received 15 November 2007; received in revised form 4 February 2008; accepted 4 February 2008
Available online 12 February 2008
Abstract
A reinvestigation of the reactions leading to arsonolipids (2,3-diacyloxypropylarsonic acids) has been carried out in order to understand why the
yields of their preparation were only moderate, although they are better than those reported for 2,3-diacyloxypropylphosphonic acid (phosphotidic
acid). Thus, the reaction of glycidol and of 3-chloro-1,2-propanediol with alkaline sodium arsenite, “Na3AsO3”, gives the desired product, 2,3-
dihydroxypropylarsonic acid, and ∼10% of an arsenic-containing glycerol dimer which is removed during the preparation of these arsonolipids.
The step which is mainly responsible for the diminished yields is due to the reaction of the –As(SPh)2 or –AsO3H− precursor with the activated acid
chlorides or carboxylic acid anhydrides to give an intermediate which cyclizes with the primary hydroxy group of the 2,3-dihydroxypropyl moiety.
This cyclization does not allow the primary hydroxy group to be acylated. Such cyclization could not be avoided with RCOCl/py, (RCO)2O/DMAP,
or RCOOH/DCC/DMAP acylating systems.
© 2008 Elsevier Ireland Ltd. All rights reserved.
Keywords: Arsonolipids; 2,3-Diacyloxypropylarsonic acids; Glycidol (2,3-epoxypropanol); Trisodium arsenite
1. Introduction
and dithioarsonites (Eqs. (1) and (2), respectively) took place
under drastic conditions (150–160 ◦C/5–18 h), we prepared the
dithioarsonite 6, acylated it (RCOCl/py) to 7 and after hydrogen
peroxide oxidation of 7 the arsonolipids 5 were prepared with
essentially thesameyields(20–50%) (Servesetal., 1993). Better
A new class of lipids 5 (see Scheme 1 for formulae), called
arsonolipids (Tsivgoulis et al., 1991a) was prepared from rac-,
R-, and S-2,3-dihydroxypropylarsonic acids 2 (Tsivgoulis et al.,
1991b; Serves et al., 1992; Lacoste et al., 1992). Thus, acylation
of the mono n-tetrabutylammonium 3 (Tsivgoulis et al., 1991b)
or the di n-tetrabutylammonium 4 (Serves et al., 1992) salts
of 2 with (RCO)2O/py and RCOCl/py, respectively, gave the
arsonolipids 5 in only 20–40% yields.
R–As(SRꢀ)2 + 2RꢀꢀCOCl → R–AsCl2 + 2RꢀꢀCOSRꢀ
(1)
R–As(SRꢀ)2 + 2(RꢀꢀCO)2O → R–As(OCORꢀꢀ)2+ 2RꢀꢀCOSRꢀ
(2)
Based on the report (Kamai and Chadaeva, 1957) that reac-
tion between acyl chlorides or carboxylic acid anhydrides
yields (60–75%) were obtained for the preparation of arsi-
nolipids 10 from 11 by a procedure analogous to that
shown in Scheme 1 (11 in place of 6) using the RCOCl/py
and RCOOH/DCC/DMAP acylating systems (Kordalis and
Ioannou, 2000), implying that in the reaction of 6 with
RCOCl/py some side-reactions took place and two probable by-
products were reported (Serves et al., 1993).
Abbreviations: RCOOH, fatty acid; RCOCl, fatty acyl chloride; (RCO)2O,
fatty acid anhydride; RC(O)SH, thioalkanoic acid; py, pyridine; DMAP,
4-dimethylaminopyridine; DCC, N,Nꢀ-dicyclohexylcarbodiimide; DHU, 1,3-
dicyclohexylurea; Phosphotidic acid, 2,3-diacyloxypropylphosphonic acid;
Glycidol, 2,3-epoxypropanol; Capric acid, decanoic acid; Myristic acid,
decatetranoic acid; Palmitic acid, hexadecanoic acid; Thiomyristic acid, thiote-
tradecanoic acid; Thiopalmitic acid, thiohexadecanoic acid.
∗
∗∗
Corresponding author. Tel.: +30 610 996025.
Corresponding author. Tel.: +30 610 997107.
E-mail addresses: tsivgoulis@chemistry.upatras.gr (G.M. Tsivgoulis),
ioannou@chemistry.upatras.gr (P.V. Ioannou).
0009-3084/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.chemphyslip.2008.02.001

114
G.M. Tsivgoulis, P.V. Ioannou / Chemistry and Physics of Lipids 152 (2008) 113–121
Scheme 1.
In order to avoid the preparation of acyl chlorides, thus
2. Experimental
simplifying the synthesis of 5, we used the acylating system
RCOOH/DCC/DMAP (Eibl et al., 1983) which is equivalent to
the (RCO)2O/DMAP system (Gupta et al., 1977) which were
used for the synthesis of lipids and phospholipids. The acy-
lation of 6 with RCOOH/DCC/DMAP followed by oxidation
gave 5 with yields 35–55% (this work) essentially similar to
those obtained with the RCOCl/py system, but lower than those
obtained for the preparation of arsinolipids 10 (Kordalis and
Ioannou, 2000) and pseudo-arsonolipids 12 (∼80% from 2-
hydroxy-1,3-bis(arsonic acid) following the sequence: reduction
with thiophenol, acylation, and hydrogen peroxide oxidation)
(Terzis and Ioannou, 2002). The yields of arsonolipids 5 are,
however, much better than the yields (18–25%) reported for
the acylation of L-dihydroxypropylphosphonate with carboxylic
acid anhydrides in the presence of tetraethylammonium salt of
the fatty acid to give the l-phosphotidic acid (Baer and Basu,
1970).
2.1. Materials
Cholesterol 99% and capric acid were from Sigma, while
myristic acid was from Serva. N,Nꢀ-Dicyclohexylcarbodiimide
(DCC) and 4-dimethylaminopyridine (DMAP) were from
Aldrich. Racemic glycidol was prepared from 3-chloro-1,2-
propanediol(Aldrich)(RiderandHill, 1930). Palmiticanhydride
(Lapidot et al., 1969) and triphenyl trithioarsenite, As(SPh)3
(Serves et al., 1995b) were prepared by literature procedures.
The cation exchange resin Dowex AG 50W-X8 (H+) and the
anion exchange resin Dowex AG 1-X8 (acetate) were from Bio-
Rad, while silica gel 60 for column chromatography and silica
gel 60 H for thin layer chromatography (TLC) were from Merck.
EthanolanddichlomethaneweredriedoverA4 molecularsieves.
2.2. Methods
These results imply that the acylating systems RCOCl/py
and RCOOH/DCC/DMAP react with the dithioarsonite 6 at the
expected hydroxy groups and at As(III) as well, and understand-
ing the nature of the various intermediates will probably improve
the yields in the preparation of 5, since these arsonolipids have
quite interesting biophysical (Faturos et al., 2001; Gortzi et al.,
2001; Fatouros et al., 2005a,b), biochemical (Supuran et al.,
1996; Rogers et al., 1996), and biological (Gortzi et al., 2002,
2003; Antimisiaris et al., 2003, 2005; Fatouros et al., 2006)
properties.
The complete route to 5 examined in this paper (Scheme 1)
requires the synthesis of 2, its reduction by thiophenol to 6, acy-
lation of 6 and oxidation of the acylated dithioarsonite 6 to 5. In
thispaperweexaminedmorecloselythepreparationof2andof5
at large scale and compare the yields of 5 with the yields obtained
for 10 and 12. In order to understand the acylation of 6, we did
competitive experiments using cholesterol, Chol–OH, (having a
secondary –OH group which is acylated slower than a primary
–OH) in the presence of triphenyl trithioarsenite, As(SPh)3, as
a model for 6.
TLCs were run on microslides. The spots were made vis-
ible by iodine vapors (especially for “As2O3”, As(SPh)3 and
PhSSPh) followed by spraying with 35% sulfuric acid and char-
ring. Cholesterol and cholesteryl palmitate gave a red color
before being charred. IR spectra were taken in KBr discs
on a PerkinElmer, model 16PC FT-IR spectrometer, while
1H NMR (at 400 MHz) and 13C NMR (at 100 MHz) spectra
were obtained on a Bruker, model DPX Avance, spectrometer.
The reference standard in D2O was DSS (2,2-dimethyl-2-
silapentane-5-sulfonate, sodium salt). Elemental analyses were
done by the Centre of Instrumental Analyses, University of
Patras, Patras, Greece, and C.N.R.S., Vernaison, France.
2.3. The reaction of rac-glycidol 1 with alkaline sodium
arsenite, Na3AsO3: preparation of
rac-2,3-dihydroxypropylarsonic acid, rac-2
To a solution of arsenic(III) oxide (3.960 g, 20 mmol As2O3)
and sodium hydroxide (4.80 g, 120 mmol) in water (7.5 ml) in

G.M. Tsivgoulis, P.V. Ioannou / Chemistry and Physics of Lipids 152 (2008) 113–121
115
a 25 ml round-bottomed flask having a short and thick mag-
2.3.2. Characterization of compound 2
netic follower, was added drop-wise during 3 h racemic glycidol
(2.963 g, 40 mmol). The very viscous solution was stirred for
another 2 h, cooled (ice-water) and concentrated hydrochloric
acid (10.0 ml, 120 mmol) was added portion-wise, the pH at
the end being ∼2. The partly precipitated NaCl and As2O3,
which did not react, were centrifuged and washed with cold
water (2.5 ml). The combined supernatants contained by TLC
(MeOH/conc. NH3 4:1) the acid 2 (Rf 0.45) and As2O3 (Rf
0.50) while glycerol was not visible at Rf 0.80. The supernatants,
diluted with water (20 ml) were applied onto the strongly cation
exchange resin (Dowex AG 50W-X8, H+ form, 2.2 cm × 26 cm,
capacity 170 mmol H+) and eluted with water (500 ml) (Lacoste
et al., 1992). The eluate was evaporated (rotary, 60 ◦C) to
an oil which was dried very well (for total removal of HCl)
in vacuum to give a foam plus glass (∼6.3 g). This mass
was dissolved in water (15 ml), applied portion-wise onto the
strongly anion exchange resin (Dowex AG 1-X8, acetate form,
3 cm × 32 cm, capacity 313 mmol AcO−) and eluted with a)
water (200 ml), and b) 1 M aqueous acetic acid (1000 ml) at a
flow rate of 5 ml/min, collecting 50 ml fractions. Fractions 1–6
gave a solid plus oil (1.16 g) containing, by TLC, glycerol and
As2O3, from which only 430 mg As2O3 could be isolated by
adding water (3 ml) and then methanol (5 ml) and centrifuging.
Fractions 9–10 gave 98 mg of a glass (compound 8) and frac-
tions 11–16 gave the product 2 (4.22 g, 53%) as a glass plus
foam.
The glassy product 2 was converted into a foam by dissolving
in warm methanol (7 ml), evaporating and drying in vacuum. Its
1
m.p., IR, H NMR, and 13C NMR spectra were as published
(Tsivgoulis et al., 1991a). It was 91% pure (impurities: acetic
acid, methanol and traces of 8). For analytical purposes a sam-
ple of it (110 mg, 0.55 mmol) in a centrifuge tube was dissolved
in methanol (2 ml) and treated with a clear solution of recrys-
tallized barium hydroxide octahydrate (173 mg, 0.55 mmol) in
methanol (5 ml). After 3 h at r.t. centrifugation and drying in vac-
uum gave the barium salt of 2 (192 mg, 94%) as the dihydrate,
insoluble in methanol and in water, and decomposing slowly at
>195 ◦C (lit. Tsivgoulis et al., 1991a: 200 ◦C dec.). Calculated
for C3H7O5AsBa·2H2O (Mr 371.38): C 9.70, H 2.99%; found C
9.89, H 2.72%. IR (KBr): 3380 s, broad, 2925 m, 1664 w, 1560
m, 1414 m, 1347 w, 1072 ms, 1035 ms, 814 vs, 646 w.
When 15 mol% excess glycidol was used then the crude reac-
1
tion mixture contained by H NMR 82% 2, 9% 8, and 9%
glycerol. The isolated yield of 2 was ∼60%.
When 3-chloro-1,2-propanediol was used as substrate
(Tsivgoulis et al., 1991a) the crude reaction mixture contained,
by 1H NMR, 65% 2, 8% 8, and 27% glycerol.
2.4. Acylation of cholesterol in the presence of an
equimolar quantity of triphenyl trithioarsenite, As(SPh)3
2.4.1. General procedure
Triphenyl trithioarsenite (20 mg, 0.05 mmol), cholesterol
(19.4 mg, 0.05 mmol), palmitic anhydride (x·0.05 mmol), and
DMAP (x·0.05 mmol or x·0.005 mmol) (x = 1–6) were dissolved
in dry dicholoromethane (2 ml) and stirred as shown in Table 1.
In all cases TLC (petroleum ether, Et2O/petroleum ether 1:10
and 1:3) showed (cholesterol which did not “react” even in the
presence of excess palmitic anhydride), cholesteryl palmitate,
phenyl thiopalmitate, and diphenyl disulfide. Evaporation and
drying in vacuum gave white solids that were examined by IR:
acid anhydride was detected only in the cases with large excess
of palmitic anhydride (entries 4–7, 10,11, Table 1). Then, the
solid of each run was dissolved in cholorform (1 ml) and chro-
matographed (silica gel 15 g in petroleum ether) eluting with
(a) petroleum ether (25 ml) and (b) ether/petroleum ether 1:10
(100 ml). The fatty acid produced by hydrolysis of the anhydride
was retained on the column, as was the free cholesterol. The first
80–90 ml gave ∼40 mg of a solid which was analysed by IR
and 1H NMR. The IR spectra showed the absence of anhydride
(1802 cm−1), thepresenceofcholesterylpalmitate(1740 cm−1),
phenyl thiopalmitate (1704 cm−1) and diphenyl disulfide (1576,
740, and 688 cm−1). The 1H NMR spectra permitted the iden-
tification (Tsivgoulis and Ioannou, submitted for publication)
and quantitation of As(SPh)3 (7.44–7.47 ppm for H-2 and H-6
of PhS–), cholesteryl palmitate (4.60 ppm RCOOCH), phenyl
thiopalmitate (2.64 ppm, triplet, for CH2CO– and 7.41, singlet,
for PhS–), and PhSSPh (7.48–7.52 ppm, apparent doublet, for
ortho protons of PhSSPh). Results are shown in Table 1.
The solids (343 mg) from several runs containing cholesteryl
palmitate, PhSSPh, and RCOSPh, were pooled and recrystal-
lized three times from acetone, to give pure, by TLC, cholesteryl
2.3.1. Characterization of compound 8
By triturating the glassy compound 8 with MeOH/Et2O and
evaporation, it was transformed into foam, keeping tenaciously
methanol and/or acetic acid, and having Rf 0.42, while 2 has
Rf 0.31. It decomposed at ∼213 ◦C. IR (KBr) (all bands are
somewhat broad): 3412 vs, 2924 m, 1636 m, 1460 w, 1397 w,
1
1343 w, 1257 w, 1222 w, 1124 s, 1044 s, 902 s, 764 m. H
NMR (D2O, DSS), δ = 2.85 (m, 2 H, H-1), 3.65 (m, 6 H, H-3,
H-4, H-6), 3.90 (m, 1 H, H-5), 4.35 (m, 1 H, H-2). 13C NMR
(D2O, DSS), δ = 40.96 C-1, 65.22 C-6, 67.15 C-2, 73.13 C-4,
74.82 C-5, 77.18 C-3. The 13C peaks were assigned based on
the spectra of polymerized glycidol (Sunder et al., 1999; Tokar et
al., 1994; Dworak et al., 1995; Cassel et al., 2001). For analytical
purposes, the acid 8 was converted into its barium salt as follows:
To a solution of 8 (98 mg, 0.35 mmol) in methanol (2 ml) and
water(0.5 ml)aclearsolutionofrecrystallizedbariumhydroxide
octahydrate (113 mg, 0.35 mmol) in methanol (5 ml) was added
and the solution was evaporated and dried in vacuum. The solid
was suspended in boiling methanol (3 ml), cooled to r.t. and
centrifuged to give the barium salt of 8 (117 mg, 82%). It is
soluble in water and insoluble in warm methanol and in warm
acetone. M.p. ∼205 ◦C dec. Calculated for C6H13O7AsBa (Mr
409.42): C 17.60, H 3.20%; found: C 17.64, H 3.15%. IR (KBr)
(all bands are somewhat broad): 3412 s, 2920 m, 1460 m, 1452
mw, 1320 mw, 1111 ms, 1062 ms, 818 vs. 1H NMR (D2O, DSS),
δ = 2.05 (m, 2 H, H-1, 3.57 and 3.65 (m, 6 H, H-3, H-4, H-6),
3.92 (m, 1 H, H-5), 4.26 (m, 1 H, H-2). 13C NMR (D2O, DSS),
δ = 39.63 C-1, 65.30 C-6, 68.81 C-2, 73.04 and 73.14 C-4, 74.74
and 74.69 C-5, 77.72 and 77.75 C-3.

116
G.M. Tsivgoulis, P.V. Ioannou / Chemistry and Physics of Lipids 152 (2008) 113–121
Table 1
Competitive acylation of cholesterol and of triphenyl trithioarsenite, (PhS)3As, by 4-dimethylaminopyridine-activated palmitic anhydride
Entry
Reactants
Reaction conditionsa,b
Recovered
As(SPh)3 (%)
Yields of non-arsenic-containing products
c
Chol–OHd
(mmol)
(PhS)3As
(mmol)
(RCO)2O
(mmol)
DMAPe
(mmol)
RCOSPhf,g
(%)
PhSSPhg,h
(%)
RCOOChol
(%)
1
2
0.05
0.05
0.05
0.05
0.05
0.10
0.005
0.01
r.t./24 h
28
25
25
46
20
20
23
25
(RCO)2O added slowly at
0 ◦C; then r.t./24 h
(RCO)2O added slowly to
the reaction mixture
having Et3N at 0 ◦C; then
r.t./24 h
3
0.05
0.05
0.10
0.01
35
43
14
10
4
5
6
7
8
9
10
11
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.10
0.20
0.25
0.30
0.05
0.10
0.20
0.25
0.01
0.02
0.025
0.03
0.05
0.10
0.20
0.25
r.t./24 h
r.t./24 h
r.t./24 h
r.t./24 h
r.t./24 h
r.t./24 h
r.t./24 h
r.t./24 h
18
20
0
23
19
0
21
24
81
94
21
23
46
63
52
54
66
48
28
61
70
60
0
4
30
24
19
13
24
23
10
3
a
Cholesterol was detected by TLC at the end of all experiments.
b
c
d
e
f
Palmitic anhydride was detected by IR only in entries 5–7 and 10,11.
Some (PhS)3As decomposes on the silica gel column giving PhSSPh and As2O3.
Chol–OH = cholesterol.
DMAP = 4-dimethylaminopyridine.
The yields have been calculated for the conversion (PhS)3As → 3 RCOSPh.
The yields have been calculated for the conversion 2 (PhS)3As → 3 PhSSPh.
The yields should be indicative when (RCO)2O is not in excess.
h
g
palmitate (65 mg), m.p. 74 ◦C sinters, 76–78 ◦C melts. Lit.
(Swell and Treadwell, 1955): 75 ◦C cloudy, 80.5 ◦C melts. IR
(KBr): 2916 vs, 2850 vs, 1742 s, 1464 m, 1380 mw, 1308 w,
1283 w, 1266 w, 1242 w, 1222 w, 1196 m, 1178 ms, 1012 w,
800 w, 726 w.
2.5.2. Acylation of rac-6 to rac-7
To the oily rac-6 (assuming to contain 19.2 mmol rac-
6) was added myristic acid (10.51 g, 46.1 mmol), DMAP
(560 mg, 4.6 mmol), and dissolved by dry dichloromethane
(150 ml). Then, a solution of DCC (9.30 g, 45.1 mmol) in dry
dichloromethane (50 ml) was added drop-wise during 3 h, and
the heterogeneous (1,3-dicyclohexylurea, DHU) solution was
stirred at r.t. for 2 days. TLC (CHCl3/MeOH 20:1) showed
DMAP (Rf 0.0), DHU (Rf 0.22), myristic acid (Rf 0.30, and
a strong spot at 0.80 <Rf < 1.0 containing the product and other
non-polar by-products. The solution was evaporated and dried in
vacuo (to remove dichlomethane which solubilizes some DHU)
to give 27.3 g of a solid. This was suspended in ether (100 ml)
filtered through celite and washed with ether (3 × 50 ml) to give
a clear, faint yellowish solution.
2.5. Preparation of rac-5 (R = C13H27) via
RCOOH/DCC/DMAP acylation of rac-6
2.5.1. Reduction of rac-2 to rac-6 with thiophenol
To a solution of rac-2 (3.844 g, 19.2 mmol) (containing 3%
methanol and 3% acetic acid) in methanol (100 ml), was added
portion-wise thiophenol (8.21 ml, 76.8 mmol) and the solution
with the precipitated diphenyl disulfide stirred overnight. Evap-
oration and drying in vacuo gave a solid which was extracted
with boiling petroleum ether (2 × 50 ml and 2 × 25 ml) decant-
ing while warm. The dried product rac-6 contaminated by 5%
diphenyl disulfide was a semi-opalescent oil (7.14 g, expected
7.07 g). IR (neat): 3354 br, s, 3056 w, 2922 w, 2870 w, 1580 m,
1476 s, 1438 s, 1398 w, 1082 s, 1060 s, 1024 s, 1002 w, 960
w, 868 w, 742 vs, 690 vs. 1H NMR (CDCl3), δ = 2.29 (m, 2 H,
CH2As), 3.55 and 3.70 (both doublet of doublets, 2 H, CH2OH),
4.22 (m, 1 H, CHOH), 7.22 (d, J = 0.8 Hz, 6 H, meta and para
hydrogens), 7.40 (m, 4 H, ortho hydrogens) and protons of the
phenyl disulfide impurity. 13C NMR (CDCl3), δ = 35.83 CH2As,
67.46 CH2OH, 69.72 CHOH, 127.59 para-C, 129.03 ortho-C,
133.10 and 133.39 ipso-C, 133.93 meta-C. From the diphenyl
disulfide impurity: the ipso-C was not visible at 136.97, the ortho
and meta-C coincided with those of 3 and a very small peak at
127.14 (para-C) was visible.
2.5.3. Oxidation of rac-7 to rac-5 and PhSSPh with H2O2
Totheethersolutionofimpurerac-7wasaddedwater(20 ml),
stirred vigorously to give an emulsion, and then aqueous hydro-
gen peroxide 30% (5.23 ml, 46.1 mmol) was added drop-wise
(10 min). The emulsion gave a suspension which after 3 h was
centrifuged. The upper ether phase was decanted and worked-up
(see below), and the water (lower phase, positive for hydrogen
peroxide by the TiOSO4 test) was syringed off. The middle solid
phase was the product rac-5 (R = C13H27), which was dried in
vacuum to give 7.58 g (expected 11.90 g) of a white solid. It was
recrystallized from boiling absolute ethanol (10 ml/g), cooling at
r.t. overnight. The product rac-5 (R = C13H27) (6.239 g, 52.4%),
pure by TLC (CHCl3/MeOH 20:1, Rf 0.08, CHCl3/CH3COOH
10:1, Rf 0.24), shrinks at 85 ◦C and melts at 90–92 ◦C. (lit.

G.M. Tsivgoulis, P.V. Ioannou / Chemistry and Physics of Lipids 152 (2008) 113–121
117
Tsivgoulis et al., 1991b: 91–93 ◦C). Its IR (KBr) spectrum is
as published (Tsivgoulis et al., 1991b).
1H NMR (CDCl3), δ = 0.88 (t, 6 H, H-10 and H-10ꢀ), 1.26
(m, 24 H, H-4 to H-9 and H-4ꢀ to H-9ꢀ;), 1.57 (m, 4 H) H-3 and
H-3ꢀ, 2.26 (m, 4 H), H-2 and H-2ꢀ), 2.70 (m, 2 H), H-1ꢀꢀ, 4.12
and 4.35 (both doublet of doublets, 2 H, H-3ꢀꢀ), 5.41 (m, 1 H,
H-2ꢀꢀ), 7.49 (m, 2 H, meta-H), 7.56 (m, 1 H, para-H), 7.80 (d,
J = 7.2 Hz, 2 H, ortho-H), 8.1 and 8.3 (broad, 2 H, H2O). 13C
NMR (CDCl3), δ = 14.11 C-10 and C-10ꢀ, 22.70 C-9 and C-9ꢀ,
24.87 C-3, 24.94, C-3ꢀ, 29.08 – 29.47 C-4 to C-7 and C-4ꢀ to
C-7ꢀ, 31.91 C-8 and C-8ꢀ, 33.82 and 34.04 C-2 and C-2ꢀ, 34.90
C-1ꢀꢀ, 64.50 C-3ꢀꢀ, 66.35 C-2ꢀꢀ, 129.37 ortho-C, 130.22 meta-C,
132.73 para-C, 134.29 ipso-C, 172.40 C-1ꢀ, 173.12 C-1.
1H NMR (CDCl3), δ = 0.88 (t, 6 H, H-14 and H-14ꢀ), 1.26
(m, 40 H, H-4 to H-13 and H-4ꢀ to H-13ꢀ), 1.61 (m, 4 H, H-
3 and H-3ꢀ), 2.33 (m, 4 H, H-2 and H-2ꢀ), 2.71 (d, J = 6.8 Hz,
2 H, H-1ꢀꢀ), 4.19 and 4.37 (both doublet of doublets, 2 H, H-
3ꢀꢀ), 5.47 (m, 1 H, H-2ꢀꢀ), 6.34 (broad, 2 H, As(OH)2). 13C NMR
(CDCl3), δ = 14.13 C-14 and C-14ꢀ, 22.71 C-13 and C-13ꢀ, 24.66
C-3, 24.83, C-3ꢀ, 29.13–29.70 C-4 to C-11 and C-4ꢀ to C-11ꢀ,
31.93 C-12 and C-12ꢀ, 33.19 C-1ꢀꢀ, 33.97 C-2 and C-2ꢀ, 63.97
C-3ꢀꢀ, 65.21 C-2ꢀꢀ, 172.49 C-1ꢀ, 173.09 C-1. The assignments are
based on the work of Amato et al. (1990).
3. Results and discussion
3.1. The reaction of alkaline arsenite with electrophiles
The reaction of alkaline sodium arsenite, “Na3AsO3”, with
short chain alkyl halides is the best method (the Meyer reaction;
Meyer, 1883) for the preparation of aliphatic arsonic acids. With
medium chain alkyl halides the reaction is very “temperamental”
(Pietsch, 1965) and with long chain alkyl halides it does not
work because the mixing of a lipophilic halide with the aqueous
arsenite is very poor (Dixon, 1997). The lipophilicity of the
medium and long chain alkyl halides can be overcome by using
as substrates ␣-halo acids RCHXCOOH which in the Meyer
reaction will give RCH(AsO3H2)COOH that can be thermally
decarboxylated to RCH2AsO3H2 (Adams et al., 1984; Ioannou,
2002a). Alkaline arsenite reacts with epoxides (Chelintsev and
Kuskov, 1946) giving ␣-hydroxy arsonic acids and we exploited
this feature in order to prepare arsonolipids 5 (Serves et al., 1992)
and arsinolipids like 10 and more complex ones (Kordalis and
Ioannou, 2000; Ioannou, 2002) using glycidol as a substrate.
The nucleophile in the Meyer reaction is the trianion,
AsO33−, which is present in minute amounts in aqueous
solutions of sodium arsenite (Serves et al., 1994). Its con-
centration can increase in the presence of excess sodium
hydroxide (an undesirable strategy because the HO− is also
nucleophilic) or by using as concentrated solutions of sodium
arsenite as possible. The reaction of glycidol with concentrated
sodium arsenite gives a very viscous solution of disodium 2,3-
dihydroxypropylarsonate (see 4) and stirring is very difficult
when the preparation is run at a large scale [e.g. 40 or 80 mmol
As(III)].
For isolation of 2 the solution is “neutralized” by an equiv-
alent amount of hydrochloric acid. When the preparation is run
at a small scale [2–10 mmol As(III)], evaporation and drying in
vacuum is relatively easy and extraction with absolute ethanol
gives quite pure 2 (Tsivgoulis et al., 1991a; Serves et al., 1992,
1993) free of 8 but contaminated with As2O3, ethanol and, occa-
sionally, with traces of glycerol. When the synthesis is run on
a 40 or 80 mmol As(III) scale, the drying is not convenient and
therefore the sodium chloride is removed by a strongly cation
exchange resin (Lacoste et al., 1992) and the eluate, after evapo-
ration, is dried very well in order to remove all hydrochloric acid
becauseitdiminishesthecapacityofthestronglyanionexchange
resin which is used to purify 2. From the latter resin, elution with
water removes As2O3, glycerol (Serves et al., 1995a) and any
The ether phase was evaporated to give a solid (10.50 g)
in which diphenyl disulfide (4.19 g) and excess myristic acid
(1.75 g) should be present. It was chromatographed (silica gel
70 g in ether) eluting with a) ether (350 ml), b) CHCl3/MeOH
10:1 (200 ml), and c) CHCl3/MeOH 5:1 (300 ml), collecting
50 ml fractions. Fractions 1–7 gave 6.41 g of a mixture (by
1H NMR and TLC) of diphenyl disulfide (∼3 g), myristic acid
(∼2.5 g), and phenyl thiomyristate (RCOSPh) (∼1 g). Fractions
8–9 gave an oil (96 mg) which could not be identified. Fractions
1
10–13 gave a foam (2.29 g) which, by IR and H NMR, con-
tained RCOO– and –AsO3H2 groups, melted at 48–50 ◦C but
on recrystallization from methanol (8 ml) no chromatographi-
cally pure compound could be isolated. It is probably the lyso
compound which underwent acyl group migration (Serdarevich,
1967) during the work-up. Fraction 14–17 gave a solid (1.15 g)
containing, by IR and 1H NMR, RCOO– and –AsO3H2 groups
and DMAP.
2.6. Preparation of the arsinolipid rac-10 (R = C9H19) via
RCOOH/DCC/DMAP acylation of rac-11
This was prepared according to Kordalis and Ioannou (2000),
in dry dichloromethane. After reduction, acylation and oxi-
dation, the product did not precipitate and it was isolated by
column chromatography (silica gel 50 g in ether for a scale of
2.2 mmol). Elution with a) Et2O (400 ml) b) CHCl3/MeOH 1:1
(10 ml, in order to push the product into the silica gel), and c)
CHCl3/MeOH 10:1 (800 ml) collecting 100 ml fractions gave
the product rac-10 (R = C9H19) in the fractions 10–11 as an
oil (905 mg, 70%). The product absorbs moisture very easily
(Kordalis and Ioannou, 2000). Calculated for C29H49O6As·H2O
(Mr 586.62): C 59.37, H 8.76%; found 59.80, 8.83%. IR (neat):
1740 vs (C O), 892 m (As O), 756 s (As–OH).

118
G.M. Tsivgoulis, P.V. Ioannou / Chemistry and Physics of Lipids 152 (2008) 113–121
base catalysed polymerized glycidol (Sunder et al., 1999). From
this fraction only 25–30% of the As2O3 that did not react was
recovered because As2O3 is volatile in water when evaporated
(rotary, 60 ◦C). Elution with 1 M aqueous acetic acid gave small
amounts (∼10%) of 8 as a pair of diastereoisomers, running just
above 2 on TLC and its detection in the original mixture was not
possible. Next, 2 was eluted either free of 8 or containing traces
of it.
Condensation by-products, analogous to 8, have also been
encountered in the reaction of alkaline arsenite with dl-1,3-
butadiene diepoxide (Tsivgoulis et al., 2007).
3.2. The competitive reaction of DMAP-activated
carboxylic acid anhydride with cholesterol and triphenyl
trithioarsenite (PhS)3As
When a concentrated solution of aqueous alkaline sodium
arsenite reacted with rac-3-chloro-1,2-propanediol, the rela-
tive molecular proportions of 2, 8 and glycerol in the crude
reaction mixture were 1:0.09:0.89 (Tsivgoulis et al., 1991a),
when reacted with stoichometric amount of rac-glycidol the
In order to understand the origin of the moderate yields of 5
obtained by RCOOH/DCC/DMAP acylation of 6 in dry, non de-
aerated dichloromethane, we run competitive experiments using
equimolar quantities of cholesterol and (PhS)3As according to
Eq. (3) under various acylating conditions, Table 1.
(3)
proportions were 1:0.11:0.27 (this work) and when reacted with
15 mol% excess rac-glycidol the proportions were 1:0.08:0.24
(this work). The yields of 2 from rac-glycidol (60–80%) (Serves
et al., 1992; this work) are better than the yields from rac-
3-chloro-1,2-propanediol (45–50%) (Tsivgoulis et al., 1991a;
this work). The latter substrate gives more glycerol because of
the concomitant glycidol formation (Serves et al., 1994) which
consumes HO−. Lowering the OH− concentration results in a
Dioxygen in air has no effect on the acylation of cholesterol,
but pure As(SPh)3 in chloroform is air oxidized to As2O3 and
PhSSPh at an extent of 5–10% in 3 days (Haikou and Ioannou,
2006). Therefore, a small amount of PhSSPh can come from the
air oxidation of (PhS)3As.
The reactions of Eq. (3) can conveniently be followed by
TLC. In all cases free cholesterol was detected. Because the
anhydride hydrolyses on TLC, its presence was confirmed by
IR. The reaction mixture was chromatographed on silica gel.
During this step, any anhydride that remained was hydrolysed
to free acid which did not elute, as did cholesterol (which had
not reacted or produced by hydrolysis of an intermediate; see
below). Also, pure As(SPh)3 was mainly air oxidized by ∼65%
during the chromatography (Tsivgoulis and Ioannou, submitted
for publication). Therefore, the yields of As(SPh)3, reported in
Table 1, are underestimated and those of PhSSPh are overesti-
mated, but these do not affect the conclusions presented below.
The detection of free cholesterol on TLC even when large
excess of anhydride is present (entries 5–7, 10, 11; Table 1)
means that, in solution, it was somehow protected from acylation
as a compound which was hydrolytically labile. We hypothe-
size that cholesterol is “protected” as 14 by reacting with an
arsenic(III) compound, e.g. (PhS)2As–OCOR 13 (see below),
Eq. (4), because it is known (Gattow and Schwank, 1971) that
(AcO)3As reacts with alcohols giving (RO)3As and AcOH.
3−
decrease of the concentration of the active nucleophile AsO3
(Serves et al., 1994), and glycidol is hydrolysed to glycerol under
HO− catalysis.
The arsonic acid 2 could not be obtained pure. When isolated
by extraction and “crystallization”, it contained small amounts
of As2O3, ethanol, traces of glycerol, and, probably, water
(Tsivgoulis et al., 1991a; Serves et al., 1992, 1993). Isolation
by anion exchange chromatography, however, gives 2 contami-
nated by acetic acid, methanol, traces of 8, and, probably, water
(Section 2.3). The tenaciously held acetic acid and methanol
are efficiently removed during the preparation of 6. The other
impurity 8, is removed during “crystallization” of 2 (Tsivgoulis
et al., 1991a; Serves et al., 1992) or during the preparation of
the dilithium (Lacoste et al., 1992) or barium (this work) salts of
2. If not removed during the preparation of 6 it can be removed
from the arsonolipid 5 due to different solubility of acylated 8
(this work).
The assignment of the structure of the by-product 8 (instead
of 9) is based on the 13C NMR spectra of polymerized glycidol
in CD3OD (Sunder et al., 1999) and in D2O (Tokar et al., 1994;
Dworak et al., 1995) and on the work of Cassel et al. (2001). The
production of 8, probably containing traces of 9, comes from the
primary alkoxide of the anion 4. Our data indicate that reaction
of the alkoxide of 8 with glycidol to give a trimer is unlikely.
The dimeric 8 is obtained as a pair of diastereoisomers in equal
proportions (by 13C NMR of its barium salt). The production
of 8 in essentially the same proportion compared to 2 starting
from either glycidol or 3-chloropropane-1,2-diol, implies that
the latter gives the epoxide much faster than the Meyer reaction
on –CH2Cl, a conclusion which has been reached earlier (Serves
et al., 1994).
(PhS)2As–OCOR + Chol–OH → RCOOH+(PhS)2As–OChol
13
14
H2O
−→ (PhS)2As–OH + Chol–OH
(4)
Then, 14 hydrolyses on the silica gel giving free cholesterol,
Eq. (4), because alkoxy esters of As(III) are not hydrolytically
stable (Brill and Campbell, 1973; Baer et al., 1983). There-
fore, the intermediate such as 13 must have been formed quite
fast; in other words the reaction of As(SPh)3 with activated acid
anhydride competes with the acylation of cholesterol.
The yields of the wanted product, cholesteryl palmitate,
reaches a plateau (60–70%) when x = ∼3, Eq. (3), irrespec-
tive of the amount of DMAP (entries 4–7 and 9–11, Table 1).

G.M. Tsivgoulis, P.V. Ioannou / Chemistry and Physics of Lipids 152 (2008) 113–121
119
Scheme 2.
The remaining cholesterol is “blocked”, as described above.
Therefore, a slight excess of anhydride with catalytic amount
of DMAP gives the best yields of cholesteryl ester without
inactivation of the secondary –OH does not take place
intramolecularly because of the strain of a 4-membered ring
would have possessed if it had been formed.
wasting anhydride. The data of Table 1 (entries 2 and 3) show
that at low temperatures the As(SPh)3 is more reactive towards
DMAP-activated anhydride than the secondary –OH group, giv-
ing ∼45% thiol ester.
The moderate yields (∼55%) in the preparation of 5 by
RCOOH/DCC/DMAP acylation of 6 can be understood by the
formation of a 5-membered ring, 17 from 18. Evidence for
the formation of 17 is the detection of phenyl thiomyristate,
RC(O)SPh, arising via reactions analogous to those shown in
Scheme 2. Acylation of 17 and hydrogen peroxide oxidation
should give a lyso-arsonolipid which is removed during the iso-
lation of 5. The peroxide oxidation can convert any >As–SPh and
> As–COR into As(V) because it is known that compounds of
As(III), e.g. Ar–As(SR)2 are oxidized by I2/H2O to arsonic acids
Ar–AsO3H2 (Barber, 1929) and compounds such as Ph2As-
COMe (Steinkopf et al., 1928) or Me2As-COMe (Albers et al.,
1952) are oxidized to arsinic acids Ph2AsO2H or Me2AsO2H by
O2/H2O and warming or spontaneously, respectively. From the
preparation of 5 (R = C13H27) we isolated a significant amount
of a low melting foam which most likely is lyso-arsonolipid.
However, we could not isolate a chromatographically pure sam-
ple probably because acyl migration (Serdarevich, 1967) took
place during the attempted purification.
The nature of the arsenic-containing compounds formed in
these competitive experiments is not certain because of their
lability and reactivity, but we can offer an approach to them.
Previous results showed that when 1 mol As(SPh)3 reacts with
3 mol (RCO)2O in the presence of 0.3 mol DMAP then ∼60%
RCOSPh is produced, according to Scheme 2, by attack of PhS−
on the activated carbonyl carbon of As+–COR 15, in a man-
ner analogous to DMAP+–COR, while attack at PhS– of 15 to
give PhSSPh and 16 is not favored (Tsivgoulis and Ioannou,
submitted for publication).
Compound 13, being extremely reactive (Gattow and
Schwank, 1971), reacts with cholesterol as per Eq. (4), giving
(PhS)2As–OChol 14, thus protecting cholesterol from acylation.
It should be noted that 14 could further react with the activated
anhydride and cholesterol as per Scheme 2 and Eq. (4).
The better yields (∼70%) obtained for arsinolipids 10 are
probably due to a slower nucleophilic attack of As(III) of 11 on
the C O carbon of the activated carboxylic acid anhydride com-
pared with As(III) of 6, and/or to a smaller extent of cyclization
of 20 to 19 thereby increasing the amount of the primary –OH
acylation.
3.3. The preparation of rac-5 (R = C13H27) and rac-10
(R = C9H19) via RCOOH/DCC/DMAP acylation of rac-6
and rac-11, respectively, and spectral properties of the
products
Analogous cyclization may have taken place when we
(Tsivgoulis et al., 1991a; Serves et al., 1992) acylated salts of 2,
The above arguments can explain the high yields (∼80%)
in the preparation of 12 (Terzis and Ioannou, 2002) because

120
G.M. Tsivgoulis, P.V. Ioannou / Chemistry and Physics of Lipids 152 (2008) 113–121
e.g. 3 (an As(V) compound), with (RCO)2O/py and RCOCl/py,
respectively. In these cases, the mixed anhydride which should
have been formed (see Mangroo and Gerber, 1988) suffered
cyclization by attack of the primary –OH on As(V) giving the
cyclic monoester of As(V), thus “protecting” the primary –OH
from acylation.
Therefore, with 23 and 24, intermediates having the
>As–OCOR group can be formed and cyclization most likely
cannot be avoided in their catalysed acylation.
Acknowledgment
The solid state IR spectra of the arsonolipids 5 (Tsivgoulis
et al., 1991b) and arsinolipids 10 (Kordalis and Ioannou, 2000)
have been discussed. The arsinolipid 10 (R = C9H19), however,
showed only one band at 756 cm−1 for ␯(As–OH) implying
that there is only one type of hydrogen-bonding. The high-
We thank the referees for their suggestions which improved
the presentation of this paper.
References
1
resolution H NMR spectra of 5 (R = C13H27) resembles that
Adams, S.R., Sparkes, M.J., Dixon, H.B.F., 1984. The arsonomethyl analogue
of adenosine 5ꢀ-phosphate. Biochem. J. 221, 829–836.
Albers, H., Ku¨nzel, W., Schuler, W., 1952. Zur Kenntnis der acylierten
Arsenwasserstoff- und Phosphorwasserstoff-Derivate und der isoarsile.
Chem. Ber. 85, 239–249.
Amato, M.E., Irgolic, K.J., Junk, T., Pappalardo, G.C., Perly, B.,
1990. 1H and 13C NMR spectra of rac-2,3-bis(palmitoyloxy)propyl(2-
trimethylarsonioethyl) phosphonate, an arsenic-containing phosphonolipid.
Magn. Reson. Chem. 28, 856–861.
Antimisiaris, S.G., Ioannou, P.V., Loiseau, P.M., 2003. In vitro antileishma-
nial and trypanocidal activities of arsonoliposomes and preliminary in vivo
distribution in BALB/c mice. J. Pharm. Pharmacol. 55, 647–652.
Antimisiaris, S.G., Klepetsanis, P., Zachariou, V., Gianopoulou, E., Ioannou,
P.V., 2005. In vivo distribution of arsenic after i.p. injection of arsonolipo-
somes in balb-c mice. Int. J. Pharm. 289, 151–158.
Baer, C.D., Edwards, J.O., Rieger, P.H., Silva, C.M., 1983. Reaction kinetics of
several alkyl arsenite hydrolyses. Inorg. Chem. 22, 1402–1404.
Baer, E., Basu, H., 1970. Synthesis of ␣-phosphotidic acids. Can. J. Biochem.
48, 1010–1013.
Barber, H.J., 1929. Some derivatives of arylthioarsinous acids. J. Chem. Soc.,
1020–1024.
Brill, T.B., Campbell, N.C., 1973. Arsenites and antimonites. II. Vibrational,
nuclear quadrupole resonance, and mass spectral properties of arsenic(III)
and antimony(III) esters and thioesters. Inorg. Chem. 12, 1884–1888.
of the arsenic-containing phosphonate studied in much detail
by Amato et al. (1990). In 5 the RCOOCH proton resonates
0.25 ppm downfield compared with the RCOOCH proton of
the phosphonate and it may be attributed to the different polar
+
groups: –CH2AsO3H2 and –CH2OP(O)(O−)CH2CH2AsMe3
.
The high-resolution 13C NMR spectrum of 10 (R = C9H19)
resembles that of 5 (R = C13H27). The 1H NMR spectrum
showed the presence of water, also found by elemental chemical
analysis, and this water absorption is a feature of arsinolipids
(Kordalis and Ioannou, 2000).
3.4. Conclusion
The experiments reported herein threw light on the step that
mostaffectsthepreparationofarsonolipids 5. Thisisthereaction
of an activated acyl group with the As(III) atom of 6. The product
obtained, 18, intramolecularly gives a cyclic compound, 17, thus
blocking the primary –OH group of the substrate from being
acylated.
Searching for As(III) substrates other than 6 for the prepa-
ration of 5, the dichloroarsine 21 and the diamide 22 deemed
unsuitable because they were difficult to prepare (Serves et
al., 1993). Compound 23 (Ioannou, 2000) should not be inert
towardsacatalysedacylationbecauseacylationattheAs(III)can
take place, since it is known that ␣-furylarsine oxide reacts with
acetic anhydride at 80–90 ◦C to give ␣-furyl diacetoxyarsine
Cassel, S., Debaig, C., Benvegnu, T., Chaimbault, P., Lafosse, M., Plusquel-
lec, D., Rollin, P., 2001. Original synthesis of linear, branched and cyclic
oligoglycerol standards. Eur. J. Org. Chem., 875–896.
Chelintsev, G.V., Kuskov, B.K., 1946. Diadic tautomerism. Zh. Obshch. Khim.
16, 1481–1484 (Chem. Abs. 41 5441a (1947)).
Dixon, H.B.F., 1997. The biochemical action of arsonic acids especially as
phosphate analogues. Adv. Inorg. Chem. 44, 191–227.
Dworak, A., Walach, W., Trzebicka, B., 1995. Cationic polymerization of gly-
cidol. Polymer structure and polymerization mechanism. Macromol. Chem.
Phys. 196, 1963–1970.
´
(Etiene, 1947), Eq. (5). Finally, esters like 24 with anhydrides
can react at As(III) as well because it is known that, e.g. glycol
ester of phenylarsonous acid and acetic anhydride at reflux for
6 h gave phenyl diacetoxyarsine (Kamai and Chadaeva, 1956),
Eq. (6).
Eibl, H., McIntyre, J.O., Fleer, E.A.M., Fleischer, S., 1983. Synthesis
of labeled phospholipids in high yield. Methods Enzym. 98, 623–
632.
´
Etiene, A., 1947. Arse´nicaux et antimonie´s furyliques-␣. II. Arse´nicaux pri-
maries, secondaires et tertiaires de´rive´s du furan non substitute´. Bull. Soc.
Chim. France, 47–50.
Faturos, D.G., Gortzi, O., Klepetsanis, P., Antimisiaris, S.G., Stuart, M.C.A.,
Brisson, A., Ioannou, P.V., 2001. Preparation and properties of arsonolipid-
containing liposomes. Chem. Phys. Lipids 109, 75–89.
−H2O
H2O
R–As(OH)2 ꢀ R–As = O(CH−C→O) OR–As(OCOCH3)
(5)
3
2
2
Fatouros, D.G., Klepetsanis, P., Ioannou, P.V., Antimisiaris, S.G., 2005a. The
effect of pH on the electrophoretic behaviour of a new class of liposomes:
arsonoliposomes. Int. J. Pharm. 288, 151–156.
Ph–As(OR)2 + 2(CH3CO)2O
→ 2CH3COOR + Ph–As(OCOCH3)2
(6)

G.M. Tsivgoulis, P.V. Ioannou / Chemistry and Physics of Lipids 152 (2008) 113–121
121
Fatouros, D.G., Piperoudi, S., Gortzi, O., Ioannou, P.V., Frederik, P.,
Antimisiaris, S.G., 2005b. Physical stability of sonicated arsonoliposomes:
effect of calcium ions. J. Pharm. Sci. 94, 46–55.
Fatouros, D.G., Ioannou, P.V., Antimisiaris, S.G., 2006. Arsonoliposomes: novel
nanosized arsenic-containing vesicles for drug delivery. J. Nanosci. Nan-
otechnol. 6, 2618–2637.
Rogers, J., Yu, B.-Z., Serves, S.V., Tsivgoulis, G.M., Sotiropoulos, D.N., Ioan-
nou, P.V., Jain, M.K., 1996. Kinetic basis for the substrate specificity during
hydrolysis of phospholipids by secreted phospholipase A2. Biochemistry 35,
9375–9384.
Serdarevich, B., 1967. Glyceride isomerizations in lipid chemistry. J. Am. Oil
Chem. Soc. 44, 381–393.
Gattow, G., Schwank, H., 1971. Untersuchungen u¨ber Acetate von Elementen
der V. Hauptgruppe des Periodensystems. Z. Anorg. Allg. Chem. 382, 49–60.
Gortzi, O., Klepetsanis, P., Antimisiaris, S.G., Ioannou, P.V., 2001. Capability
of arsonolipids to transport divalent cations: a Pressman cell study. Chem.
Phys. Lipids 112, 21–29.
Gortzi, O., Papadimitriou, E., Kontoyannis, C.G., Antimisiaris, S.G., Ioannou,
P.V., 2002. Arsonoliposomes, a novel class of arsenic-containing liposomes:
effect of palmitoyl-arsonolipid-containing liposomes on the viability of can-
cer and normal cells in culture. Pharm. Res. 19, 79–86.
Serves, S.V., Tsivgoulis, G.M., Sotiropoulos, D.N., Ioannou, P.V., 1992. Syn-
thesis of (R)- and (S)-1,2-diacyloxypropyl-3-arsonic acids: optically active
arsonolipids. Phosphorus, Sulfur, and Silicon 71, 99–105.
Serves, S.V., Sotiropoulos, D.N., Ioannou, P.V., Jain, M.K., 1993. One pot syn-
thesis of arsonolipids via thiorsenite precursors. Phosphorus, Sulfur, and
Silicon 81, 181–190.
Serves, S.V., Sortiropoulos, D.N., Ioannou, P.V., Dixon, H.B.F., 1994. On the
mechanism of the Meyer reaction with epoxides and 2-haloalcohols as sub-
strates. Phosphorus, Sulfur, and Silicon 90, 103–109.
Gortzi, O., Antimisiaris, S.G., Klepetsanis, P., Papadimitriou, E., Ioannou, P.V.,
2003. Arsonoliposomes: effect of arsonolipid acyl chain length and vesicle
composition on their toxicity towards cancer and normal cells in culture.
Eur. J. Pharm. Sci. 18, 175–183.
Serves, S.V., Sotiropoulos, D.N., Ioannou, P.V., Mutenda, E.K., Sparkes, M.J.,
Dixon, H.B.F., 1995a. rac-3,4-Dihydroxybutylarsonic acid: a key enter-
mediate for isosteric arsonolipids. Phosphorus, Sulfur, and Silicon 101,
75–87.
Gupta, C.M., Radhakrishnan, R., Khorana, H.G., 1977. Glycerophospholipid
synthesis: improved general method and new analogs containing photoac-
tivable groups. Proc. Natl. Acad. Sci. U.S.A. 74, 4315–4319.
Haikou, M.N., Ioannou, P.V., 2006. The autoxidation of triaryl trithioarsen-
ites, (ArS)3As: evidence for binding and activation of triplet dioxygen by
arsenic(III). Phosphorus, Sulfur, and Silicon 181, 363–376.
Serves, S.V., Charalambidis, Y.C., Sotiropoulos, D.N., Ioannou, P.V., 1995b.
Reaction of arsenic(III) oxide, arsenous and arsenic acids with thiols. Phos-
phorus, Sulfur, and Silicon 105, 109–116.
Steinkopf, W., Schubart, I., Schmidt, S., 1928. Zur Kenntnis organischer Arsen-
verbindungen. XII.: Die Einwirkung von Sa¨urechoriden auf Diphenylarsin.
Chem. Ber. 61, 678–682.
Ioannou, P.V., 2000. On the direct reduction of arsonic acids to arsenoso
compounds: mechanisms and preparations. Appl. Organometal. Chem. 14,
261–272.
Sunder, A., Hanselmann, R., Frey, H., Mu¨lhaupt, R., 1999. Controlled synthesis
of hyperbranched polyglycerols by ring-opening multibranching polymer-
ization. Macromolecules 32, 4240–4246.
Ioannou, P.V., 2002. Synthesis of arsinolipids. II. A non-isosteric analogue of
fully acylated cardiolipin. Chem. Phys. Lipids 117, 7–18.
Ioannou, P.V., 2002a. On C-As bond formation: preparation of aliphatic arsonic
acids. Phosphorus, Sulfur, and Silicon 117, 1–14.
Kamai, G., Chadaeva, N.A., 1956. Action of halides and anhydride of acetic
acid on glycol esters of phenylarsonous acid. Dokl. Akad. Nauk SSSR 109,
309–311 (Chem. Abs. 51 1876i (1957)).
Kamai, G., Chadaeva, N.A., 1957. Alkyl esters of ethylthioarsinous [ethylth-
ioarsonous] acid. Proc. Acad. Sci. USSR 115, 717–719.
Kordalis, N.L., Ioannou, P.V., 2000. Syntheses of arsinolipids: non-isosteric
analogues of phospholipids. Appl. Organometal. Chem. 14, 273–280.
Lacoste, A.-M., Dumora, C., Ali, B.R.S., Neuzil, E., Dixon, H.B.F., 1992. Uti-
lization of 2-aminoethylarsonic acid in Pseudomonas aeruginosa. J. Gen.
Microbiol. 138, 1283–1287.
Supuran, C.T., Serves, S.V., Ioannou, P.V., 1996. Carbonic anhydrase inhibitors.
Part 33. Isozyme II inhibition with 2,3-dihydroxypropylarsonic acid and
arsonolipids. J. Inorg. Biochem. 62, 207–212.
Swell, L., Treadwell, C.R., 1955. Cholesterol esterase. VI. Relative specificity
and activity of pancreatic cholesterol esterase. J. Biol. Chem. 212, 141–150.
Terzis, A., Ioannou, P.V., 2002. Preparation of pseudo-arsonolipids:
2-acyloxypropane-1,3-bis(arsonic acids). The crystal structure of 2-
hydroxypropane-1,3-bis(arsonic acid). Chem. Phys. Lipids 117, 53–61.
Tokar, R., Kubisa, P., Penczek, S., Dworak, A., 1994. Cationic polymerization
of glycidol. Coexistence of the activated monomer and active chain end
mechanism. Macromolecules 27, 320–322.
Tsivgoulis, G.M., Sotiropoulos, D.N., Ioannou, P.V., 1991a. 1,2-
Dihydroxypropyl-3-arsonic acid: a key intermediate for arsonolipids.
Phosphorus, Sulfur, and Silicon 57, 189–193.
Lapidot, Y., Barzilay, I., Hajdu, J., 1969. The synthesis of diacyl-dl- (and -l-)
␣-glycerol phosphate. Chem. Phys. Lipids 3, 125–134.
Mangroo, D., Gerber, G.E., 1988. Phospholipid synthesis: effects of solvents
and catalysts on acylation. Chem. Phys. Lipids 48, 99–108.
Meyer, G., 1883. Ueber einige anomale Reaktionen. Ber. Dtsch. Chem. Ges. 16,
1439–1443.
Pietsch, R., 1965. Untersuchungen zur Umsetzung von Alkylbromiden mit Tri-
natrium arsenite. Monatsch. Chem. 96, 138–146.
Tsivgoulis, G.M., Sotiropoulos, D.N., Ioannou, P.V., 1991b. rac-1,2-
Diacyloxypropyl-3-arsonic acids: arsonolipid analogues of phosphono-
lipids. Phosphorus, Sulfur, and Silicon 63, 329–334.
Tsivgoulis, G.M., Lala, M.A., Ioannou, P.V., 2007. Preparation of dl-2,3,4-
trihydroxybutylarsonic acid and dl-2,3-dihydroxybutane-1,4-bis(arsonic
acid): starting compounds for novel arsonolipids. Chem. Phys. Lipids 148,
97–104.
Tsivgoulis, G.M., Ioannou, P.V. The nucleophilicity of M(SPh)3 (M = P, As) and
PhAs(SPh)2 towards non-activated and activated carboxylic acid anhydrides,
Phosphorus, Sulfur, and Silicon, submitted for publication.
Rider, T.H., Hill, A.J., 1930. Studies on glycidol. I. Preparation from glycerol
monochlorohydrin. J. Am. Chem. Soc. 52, 1521–1530.