Preparation of 2,3,4-trihydroxybutylarsonic acid: A starting compound for novel arsonolipids

Article abstract of DOI:10.1080/10426500701506184

Possible routes for the preparation of 2,3,4-trihydroxybutylarsonic acid, a key compound for the synthesis of novel arsonolipids, were experimentally evaluated. The best substrate was found to be 3,4-epoxybutane-1,2-diol. Its reaction with alkaline sodium arsenite, "Na3AsO3," gave the arsonic acid in 50% yield, as two pairs of diastereoisomers, each pair being a racemic mixture. Copyright Taylor & Francis Group, LLC.

Full text of DOI:10.1080/10426500701506184

This article was downloaded by: [Pennsylvania State University]
On: 03 June 2013, At: 17:30
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK
Phosphorus, Sulfur, and Silicon
and the Related Elements
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/gpss20
Preparation of 2,3,4-
Trihydroxybutylarsonic Acid: A
Starting Compound for Novel
Arsonolipids
Maria A. Lala a , Gerasimos M. Tsivgoulis a &
Panayiotis V. Ioannou a
a Department of Chemistry, University of Patras,
Patras, Greece
Published online: 03 Dec 2007.
To cite this article: Maria A. Lala , Gerasimos M. Tsivgoulis & Panayiotis V. Ioannou
(2007): Preparation of 2,3,4-Trihydroxybutylarsonic Acid: A Starting Compound for
Novel Arsonolipids, Phosphorus, Sulfur, and Silicon and the Related Elements, 182:12,
2747-2760
To link to this article: http://dx.doi.org/10.1080/10426500701506184
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-
and-conditions
This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden.
The publisher does not give any warranty express or implied or make any
representation that the contents will be complete or accurate or up to
date. The accuracy of any instructions, formulae, and drug doses should be

independently verified with primary sources. The publisher shall not be liable
for any loss, actions, claims, proceedings, demand, or costs or damages
whatsoever or howsoever caused arising directly or indirectly in connection
with or arising out of the use of this material.

Phosphorus, Sulfur, and Silicon, 182:2747–2760, 2007
Copyright © Taylor & Francis Group, LLC
ISSN: 1042-6507 print / 1563-5325 online
DOI: 10.1080/10426500701506184
Preparation of 2,3,4-Trihydroxybutylarsonic Acid:
A Starting Compound for Novel Arsonolipids
Maria A. Lala
Gerasimos M. Tsivgoulis
Panayiotis V. Ioannou
Department of Chemistry, University of Patras, Patras, Greece
Possible routes for the preparation of 2,3,4-trihydroxybutylarsonic acid, a key com-
pound for the synthesis of novel arsonolipids, were experimentally evaluated. The
best substrate was found to be 3,4-epoxybutane-1,2-diol. Its reaction with alkaline
sodium arsenite, “Na3 AsO3 ,” gave the arsonic acid in ∼50% yield, as two pairs of
diastereoisomers, each pair being a racemic mixture.
Keywords 2,3,4-Trihydroxybutylarsonic acids; 3,4-epoxy-1-butene; 3,4-epoxybutane-
1,2-diol; 3,4-butene-1,2-diol; 1,3-butadiene diepoxide; chloramine T; trichloroisocyanuric
acid
INTRODUCTION
More than 100 arsenic-containing lipids or their precursors have been
found in nature in very small amounts.1 Their role is not very well un-
derstood, but it seems that they are formed in an attempt to detoxify
an organism from the arsenic because arsenic in most of these end com-
pounds is in the form of either RRꢀ3As+ or RRꢀ2As O, which is somewhat
chemically inert. A phosphonate analogue of phosphatidylcholine has
been prepared having –AsMe+3 instead of –NMe3+.2
This laboratory has long been involved in the synthesis of arsenic-
containing lipids having a functional As(V) in the sense that it can
react under very mild conditions with biologically important thiols,
4,5
e.g., glutathione. Such lipids are the arsonolipids 1 (rac-,3 R− and
S−4,5), 2 (rac-6), and 3,7 as well as the arsinolipids 4 as analogs of bis-
phosphatidic acid8 and of cardiolipin.9
From these lipids, the arsonolipids 1 have been studied further, and
they have been found to exhibit interesting biophysical properties, e.g.,
Received 15 January 2007; accepted 11 June 2007.
Address correspondence to Panayiotis V. Ioannou, Department of Chemistry, Univer-
sity of Patras, Patras, Greece. E-mail : ioannou@chemistry.upatras.gr
2747

2748
M. A. Lala et al.
they transport divalent cations through a chloroform phase10 and form
liposomes.11 They can inhibit enzymes, e.g., carbonic anhydrase,12 and
are (poor) substrates for phospholipase A2.13 From the biological point
of view they are selectively toxic towards certain cancer cells,14,15 while
they show promising antileishmanial and trypanocidal activities.16
Most of these studies have been summarized in recent reviews.1,17 In-
corporating arsenic into arsonolipids is hoped to reduce the toxicity and
the poor absorption seen in As2O3 a drug used to treat relapsing acute
promyelotic leukemia18 or multiple myeloma.19
The properties of arsonolipids 1 made us wonder what could be the
properties of arsonolipids 5 having three instead of two fatty acyl chains.
Although the ester 6 was prepared en route to D-erythritol 4-phosphate
7,20 no long-chain phospho- or phosphonolipids have been prepared.
Since esters of arsenic(V) acids are hydrolytically extremely
unstable,21 a major challenge in the synthesis of optically active 8 is
to obtain a chiral compound suitable for the creation of a hydrolyti-
cally stable C-As bond. Such compounds should be 2-haloalcohols or
epoxides,22 and can be prepared from meso-erythritol, tartaric acids,
and ascorbic acids in multistep preparations requiring regiospecific pro-
tection of the –OH group(s) and timely deprotections.
Before embarking in such preparations, we wanted to explore the
chemistry of a bit more complex 2-haloalcohols and of epoxides in their
reaction with alkaline arsenite (the Meyer reaction23,24). In this paper,
we describe our results on the preparation of the suitable substrates 10
and 12 starting from the commercially available compounds 9 and 11, as
well as on the reaction of 13 with alkaline arsenite, as shown in Scheme
1. Our aim was to find out which compounds can give the arsonic acid
8 with maximum yield and fewer and easily separable by-products.

2749

2750
M. A. Lala et al.
RESULTS AND DISCUSSION
Attempted Preparation of the Substrates 10 and 12 from 9
and 11
The epoxide ring of glycidol (2 M solution in dry THF containing an
equimolar amount of water and 1 mol % perchloric acid) was completely
opened to give glycerol after 72 h stirring at RT or in less than 30 min at
reflux. Under the same conditions, however, equimolar quantities of DL-
butadiene diepoxide 9 and water gave traces of 10, and equal quantities
of 9 and tetritol (by TLC; AcOEt/peth 1:1) at RT. At reflux after 2 h
nearly equal quantities of 9 and tetritol were detected, while after 3
h a white translucent solid appeared which probably was polymerized
diepoxide and was not studied further.
The inability of selectively opening one epoxide ring of 9 to give 10
means that 10 reacts, at least, as fast as being formed with water to give
the tetritol. The reason for this fast ring opening is not clear. A prob-
able explanation is that of local high concentration of water molecules
around the hydrophilic 10 due to hydrogen bonding.
Since alkyl halides25 and the more hydrophilic 2-haloalcohols22,26,27
are substrates for the Meyer reaction with aqueous alkaline arsenite,
we sought the preparation of 12 from 11 by in situ generating hypochlor-
ous acid according to literature procedures. Chloramine T in the pres-
ence of sulfuric acid in acetone-water gives HOCl, which adds to alkenes
having hydrogen or alkyl substituents to give chlorohydrins.28 Sim-
ilarly, trichloroisocyanuric acid in acetone-water hydrolyses to HOCl
which adds to unfunctionalized alkenes to give chlorohydrins.29
With 11 as substrate these two reagents did not afford any 12 (TLC
analysis), probably because other reactions took place. Inspection of 11
shows that it is both an allyl alcohol and a diol. Although hypochlo-
rination of allyl alcohol by hypochlorous acid does take place (giv-
ing the Markovnikov (73%) and anti-Markovnikov (27%) product)30 it
was found that chloramine T at all pH values oxidizes allyl alcohol to
acrolein (CH2 CH CH O).21,32 Therefore, compound 12 was prepared
from 10 as described below.
Preparation of the Substrates 10 and 12
Preparation of the Epoxide 10
During the epoxidation of 11 to 1033 by m-chloroperbenzoic acid in
dichloromethane, tetritol was produced by acid catalyzed ring opening
of 10 due to water in the peracid but was not detected by TLC because
it co-precipitated with the insoluble m-chlorobenzoic acid. The prod-
ucts were extracted with minimum amounts of water, the phases being
efficiently separated only by centrifugation, and the solvent water was

Diastereoisomeric 2,3,4-Trihydroxybutylarsonic Acids
2751
removed by freeze drying. The oil obtained in ∼90–95% crude yield
contained, by 1H NMR, 60–70% 10, 20–30% tetritol, <5% 11 and <5%
m-chlorobenzoic acid, and it was stable at −20◦C for >3 months.
The epoxidation of racemic 11 will produce two pairs of diastere-
1
omers, each pair being a racemic mixture. In the H NMR spectrum,
we observed the epoxide CH protons of the pairs of diastereoisomers
2
of 10 at different shifts from which a ratio of 1:1.62 was calculated. The
13C NMR spectrum showed two peaks of unequal intensity for each
of the epoxide carbon atoms and for the CHOH group, while for the
CH2OH moiety a singlet was observed due to pairs 14/15 (threo) and
16/17 (erythro). The pair which predominates should be the 14/15 one.34
Preparation of the Chloride 12 from 10
The ring opening of styrene epoxide and butadiene monoxide
by HCl gives the chlorides with the Cl attached to the secondary
carbon atom,35,36 while with epichlorohydrin and with glycidol the
Cl− anion attacks the primary carbon atom.37 The addition of crude
10 to concentrated hydrochloric acid gave the chloride 12 as two
pairs of diastereomers. As was revealed by 13C NMR spectroscopy,
CH2(OH)CH(OH)CHClCH2OH was not formed under the above
conditions. The purification of 12 from the tetritol present in crude
12 was done by precipitation, which can be conveniently followed
by TLC. However, the product 12 was contaminated by <10% trans-
1
3,4-dihydroxytetrahydrofuran38,39 as shown by the H and especially
by the 13C NMR spectra. Its formation probably took place during
the reaction by attack of the terminal OH on the protonated 10
or its primary carbocation CH2(OH)CH(OH)CH(OH)CH+. When the
2
crude chloride 12 was distilled the distillate contained mostly 3,4-
dihydroxytetrahydrofuran. In this case the cyclization took place by
nucleophilic attack of the terminal –OH group on the CH2Cl moiety.
Summarizing, the preparation of 10 from 11 as a pair of di-
astereoisomers gives impure product. The impurities do not affect the
Meyer reaction for the preparation of 8. Therefore 10 seems to be a
good substrate, and moreover, it can be obtained in optically active
forms.34 Also the chloride 12 was not obtained pure, and although

2752
M. A. Lala et al.
the contaminants do not interfere with the Meyer reaction, its very
high viscosity will require dilution, which is not desirable.22 Another
drawback is, that the Meyer reaction with 2-haloalcohols takes place
both via the direct displacement of –Cl by AsO33− and via the attack
of AsO33− on the epoxide which is formed in the alkaline medium,22,26
thus lowering somewhat the yields.
Reaction of 13 with Aqueous Alkaline Arsenite
Butadiene monoxide, 13, is the simplest vinyl oxirane and in nucle-
ophilic reactions can give four different products depending on the soft-
ness/hardness of the nucleophile.40,41 Hard nucleophiles attack at the
epoxide carbon atoms giving 1,2 or 2,1 addition products, while soft nu-
cleophiles attack at the ethylenic CH2 carbon atom giving 1,4-conjugate
addition products or 3,4-addition products.
Aqueous alkaline arsenite “Na3AsO3” contains the non-nucleophilic
H3AsO3 and H2AsO−, the slightly nucleophilic HAsO2−, and the nucle-
3
ophilic AsO33− as well as HO−.22,26 The role of the HO−3was studied with
the reaction of 13 with aqueous NaOH, which gave >90% of 11. Column
chromatography gave a fraction containing 11 and the dimer 18. The
structure of 18 was deduced by 1H NMR: δ = 4.35 (m, 1H for CH2 =CH-
CH(OH)-CH2-), 3.97 (m, 1H for CH2 =CH-CH(O-)CH2OH) and 5.76 (m,
2H for two CH2 =CHgroups); the other peaks overlap with those of 11.
The dimer was formed by attack of the anion CH2 =CHCH(O−)CH2OH
on another molecule of 13 before being protonated by water. The forma-
tion of 11 was expected since HO− is a hard base, and it is in line with
the product (3-hydroxy-4-methoxy-1-butene) obtained from the reaction
of 13 with methanolic MeO−.42
The reaction of equimolar quantities of 13 and “Na3AsO3” at RT/24
h gave “AsO33−” (by TLC). The H NMR spectrum of the crude product
1
was very complicated. From it, traces of 11 and 18 were detected and
peaks due to CH AsO2−, CH CH , and cis and trans CH CH
2
2
were observed. Therefor3e, compounds like the salts of 19 and 20 might

Diastereoisomeric 2,3,4-Trihydroxybutylarsonic Acids
2753
have been formed by the attack of the not so hard AsO33− on C-1 and C-4
of 13. Acidification to pH 7 showed the presence of CH AsO2− (δ 1H
2
3
= 2.10 ppm) and CH AsO3H− (δ 1H = 2.75 ppm), but no pure product
2
could be isolated at this stage. Acidification to pH 2 was expected to
give the free acids 19 and 20. However, the 1H NMR spectrum did not
show the signal of CH AsO3H2 at the expected27 2.85 ppm. Evidently,
2
19 or 20, or both decomposed. H3AsO4 was not produced (by magnesia
mixture test), therefore the decomposition produced H3AsO3. A proba-
ble organic decomposition product, 2,5-dihydrofuran (b.p. 66◦C), could
not be identified by 1H NMR. Unstable arsonic acids are known in the
chemistry of arsenic,43 and 19/20 may be two of them. Therefore, buta-
diene monoxide 13, which can be obtained in optically active forms,44,45
cannot be used for the preparation of the target arsonic acid 8.
Preparation of 8 from 10 and from 12
The addition of 10 to the concentrated22 solution of alkaline arsenite
was not a problem because it was not very viscous; however, the re-
action mixture became very viscous and it was difficult to stir. After
neutralization with hydrochloric acid, the sodium chloride could not be
removed by extracting the crude 8 with boiling methanol, as was done
with 2,3-dihydroxypropylarsonic acid, DPAH2.27 Removal of sodium
chloride was effected by passing the aqueous solution through a strong
cation exchange resin.46 From the eluate, the hydrochloric acid must be
completely removed because the chloride deteriorates the strong anion
exchange resin in acetate form to be used in the next step for purifica-
47
tion of 8. From this anion resin, elution with water removed As2O3
and non arsenic containing impurities, such as tetritol. Elution with
aqueous 1 M acetic acid gave the product 8 as a glass which was slowly
and with difficulty converted into a powder by trituration. The powder
kept tenaciously small amounts of acetic acid, of methanol and, proba-
bly, of water. The yields were ∼50%, which is less than the 70% yields of
DPAH2 obtained from the reaction of glycidol with alkaline arsenite.4
Due to its high viscosity, 12 has to be added as a solution to con-
centrated aqueous alkaline arsenite. Protic solvents, like water or
methanol, will dilute the concentrated alkaline arsenite thus dimin-
ishing the concentration of the nucleophilic AsO33−
.
We have chosen
26
to add 12 as a light suspension in ether into the concentrated alkaline
arsenite at 40◦C. Then, the ether evaporates and 12 enters the aque-
ous phase. After work up the product 8 was isolated in ∼30% yield
compared to 45% yield obtained for the preparation of DPAH2 using
1-chloro-2,3-propanediol.27 Comparing the yields of 8 prepared from 10
and from 12 it is clear that in the synthesis from 12 the yield is inferior.
This is due to the fact that the reaction of 2-haloalcohols with AsO33−

2754
M. A. Lala et al.
proceeds both via direct attack of AsO33− on –CH2Cl and on the epox-
ide, which is formed under the strongly alkaline conditions used.26 This
in situ cyclization consumes HO−, which is necessary to maximize the
concentration of the nucleophilic AsO33−
.
27
Starting from racemic 11 two pairs of diastereoisomeric 8 should be
obtained, each pair consisting of a racemic mixture. The prepared 8 de-
composed at ∼218◦C, while the racemic DPAH2 decomposed at 202◦C.27
The IR and NMR spectra of 8 prepared from 10 and from 12 were iden-
tical. The 13C NMR spectra showed that the two pairs of diastereoiso-
mers have been produced in unequal amounts (∼2:1), but which pair
predominates is not known with certainty. Most likely it is the pair
arising from the 14/15 pair because it was found to predominate during
the epoxidation of racemic 11.
The IR and NMR spectra of the barium salts of 8 produced from
10 and from 12 were the same. The –AsO23− band in their solid state
1
IR spectra was at 824 cm−1 as expected. In the H NMR spectra the
CH AsO2− hydrogen atoms showed a strong signal at 2.26 ppm. The
2
3
13C NMR spectra showed two signals for each carbon atom due to the
presence of diastereoisomers with a ratio of ∼3:1.
Summarizing, from the substrates tested for the preparation of non-
optically active 8 the most promising, from the point of view of steps and
yields, is the epoxide 10. Since this compound can be obtained in opti-
cally active forms,34 it can be used for the preparation of optically active
8, especially after the result obtained, that any Payne rearrangement48
will not spoil the stereochemistry.49
EXPERIMENTAL
3,4-Epoxy-1-butene (butadiene monoxide), 1,3-butadiene diepoxide
(DL,50,51 m.p. +4◦C), 3,4-dihydroxy-1-butene (3-butene-1,2-diol), m-
chloroperbenzoic acid 77%, and trichloroisocyanuric acid were pur-
chased from Aldrich. Chloramine T (Merck) was recrystallized
from water.52 Racemic glycidol was prepared from 1-chloro-2,3-
propanediol (Aldrich) and methanolic sodium hydroxide according to
the literature.53 It was pure by TLC (CHCl3/Me2CO 3:1, Rf 0.61). The
cation exchange resin Dowex AG 50W-X8 (H+) and the anion exchange
resin Dowex AG 1-X8 (acetate) were obtained from Bio-Rad, while silica
gel 60 for column chromatography and silica gel 60 H for thin layer chro-
matography (TLC) were obtained from Merck. Tetrahydrofuran (THF)
was dried by slow percolation through activated alumina.
TLC was run on microslides. Spots were made visible by iodine va-
pors (especially for “AsO33−”) and by spraying with 35% sulfuric acid
and charring. A Labconco Freezone 4.5 instrument was used for freeze-
drying, while a Kugelrohr B-585 apparatus from Bu¨chi was used for

Diastereoisomeric 2,3,4-Trihydroxybutylarsonic Acids
2755
trap-to-trap distillation. IR spectra were obtained on a Perkin-Elmer
16PC FT-IR spectrometer. 1H NMR (at 400 MHz) and 13C NMR (at 100
MHz) spectra were obtained with a Bruker DPX Avance spectrometer.
TMS was used as standard in CDCl3, while DSS [3-(trimethylsilyl)-1-
propanesulfonic acid sodium salt] was used as standard in D2O. Ele-
mental analyses were done by the Centre of Instrumental Analyses,
University of Patras, Greece.
Attempted Preparations of 10 from 9 and of 12 from 11
To a solution of diepoxide 9 (77 µL, 86 mg, 1 mmol) in dry THF (0.5 mL),
an aqueous solution containing water (1 mmol) and perchloric acid (0.01
mmol) was added and stirred at RT for 72 h. TLC (AcOEt/peth 1:1)
showed the presence of the diepoxide (Rf 0.58), traces of 10 (Rf 0.20),
and significant amount of tetritol (Rf 0.0). When a similar solution was
refluxed for 8 h nearly equal amounts of diepoxide and tetritol were
detected by TLC.
To a solution of chloramine T trihydrate (282 mg, 1 mmol) in acetone
(0.5 mL), 3,4-dihydroxy-1-butene 11 (84 µL, 88 mg, 1 mmol) was added
followed by dropwise addition of a solution of 95.5% w/w sulfuric acid
(54 µL, 98 mg, 1 mmol) in water (0.5 mL), and the solution was stirred
at RT for 3 h. After 40 min, hypochlorous acid was not detected (1 drop
of the solution added to 1 drop of aqueous sodium iodide and 1 drop of
starch solution). TLC (CHCl3/MeOH 2:1) did not show the presence of
12 at Rf 0.72.
To a solution of trichloroisocyanuric acid (78 mg, 0.33 mmol) in ace-
tone (0.5 mL), water (0.1 mL) and 11 (84 µL, 88 mg, 1 mmol) was added
and the opalescent solution was stirred at RT for 2 h. The expected
product 12 was not detected by TLC.
Preparation of Substrates 10 and 12
Preparation of 3,4-Epoxybutane-1,2-diol (10)
To
a solution of 3-butene-1,2-diol 11 (4.400 g, 50 mmol)
in dichloromethane (10 mL) was added dropwise a solution of
m-chloroperbenzoic acid (12.33 g, purity 77%) in dichloromethane
(60 mL), and the mixture was stirred at RT for 4.5 h. During the reaction
m-chlorobenzoic acid precipitated. TLC (CHCl3/MeOH 10:1) showed
peracid + acid (Rf 0.77), 11 (Rf 0.43), and 10 (Rf 0.38). Tetritol (Rf
0.05) was not detected. Filtration (sintered glass porosity 3) removed
the precipitated m-chlorobenzoic acid plus any co-precipitated tetritol.
Concentration to ∼20 mL gave a suspension which was transferred to a
large centrifuge tube and extracted with water (5 × 5 mL), the phases
being separated by centrifugation. In the aqueous phase, the water sol-
uble tetritol was detected by TLC. The combined aqueous phases were

2756
M. A. Lala et al.
filtered (Pasteur pipette plugged with cotton) to remove precipitated
m-chlorobenzoic acid, and the clear filtrate was freeze-dried for 12–16
1
h to give an oil (4.870 g), which by H NMR contained the product 10
(70%), tetritol (26%), starting compound 11 (2%), and m-chlorobenzoic
acid (1%). The calculated yield of pure 10 was 66%. 1H NMR (D2O, DSS):
δ = 2.83 [dd, 1.24 H, HA of the one diastereoisomer (CHA HB, epoxide
ꢀ
ꢀ
ring)]; 2.86 [dd, 0.76 H, HAꢀ of the other diastereoisomer (CHA HB , epox-
ꢀ
ide ring)]; 2.93 [m, 2H, HB and HB of the two diastereoisomers (CHA HB
ꢀ
ꢀ
and CHA HB , epoxide rings)]; 3.17 (m, 2H, CH of the epoxide ring of the
two diastereoisomers); 3.53-3.79 (m, 6H, CH (OH)-CH(OH) for the two
2
diastereoisomers and the protons from tetritol). 13C NMR (D2O, DSS)
(ratio of intensities in parenthesis): δ = 44.9 and 45.2 (1.8:1) for CH2 of
the epoxide ring, 52.4 and 53.7 (1:1.8) for CH of the epoxide ring, 62.8
for CH2OH, 70.6 and 71.9 (1:1.9) for CH-OH.
A sample of crude 10 (64 % purity) was stable at -20◦C for at least 3
months, while in D2O it decomposed slowly (27% left after 40 days and
11% left after 90 days at RT).
Preparation of 4-Chlorobutane-1,2,3-triol (12)
To concentrated (12 M) hydrochloric acid (2 mL) was added dropwise
while stirring 3,4-epoxybutane-1,2-diol 10 (905 mg of 70% purity), and
the yellow solution was stirred at RT for 1 h. TLC (CHCl3/MeOH 2:1)
showed the presence of 12 (Rf 0.73) and tetritol (Rf 0.42). Evaporation
(rotary, 50◦C) gave yellow, very viscous oil which was dissolved with
slight warming in a minimum amount of methanol (∼2 mL). Addition
of diethyl ether until opalescence and cooling at +4◦C overnight pre-
cipitated a brown oil. The supernatant, still containing tetritol, was
decanted, more diethyl ether was added till opalescence (total: 40 mL
diethyl ether), and cooled at +4◦C overnight. Brown oil was precipi-
tated, and the supernatant was free of tetritol (TLC analysis). It was
evaporated and dried in vacuo to give a colorless, very viscous oil (871
mg, expected 857 mg) which according to 1H NMR contained methanol
(∼2%), 3,4-dihydroxytetrahydrofuran (∼10%) and 12 (∼85%). IR (neat):
3350 vs, broad, 2942 m, 1430 m, 1306 mw, 1074 s, 1038 s, 884 mw, 748
m, 680 m. 1H NMR (D2O, DSS): δ = 3.34 (s, MeOH impurity), 3.60–3.90
(m, 6H, all methylene and methine protons of 12). Non-resolved sig-
nals of 3,4-dihydroxytetrahydrofuran in the regions of 3.93–4.04 and
4.23–4.27 ppm. 13C NMR (D2O, DSS) (ratio of intensities in paren-
theses): For the two diastereoisomers of 12: CH2Cl: δ = 48.5 and 49.7
(1.5:1); CH2OH: δ = 65.21 and 65.25 (1.3:1); CH(OH)CH2Cl: δ = 73.5
and 73.9 (1:1.4); CH(OH)CH2OH: δ = 74.4 and 74.5 (1:1.2). For 3,4-
dihydroxytetrahydrofuran: C-1, C-4: δ = 75.5; C-2, C-3: δ = 78.9 for
trans; the cis could not be identified. Que Jr. and Gray 38 for trans: C-1,

Diastereoisomeric 2,3,4-Trihydroxybutylarsonic Acids
2757
C-4: δ = 73.0; C-2, C-3: δ = 76.4 and for cis: C-1, C-4: δ = 71.5; C-2, C-3:
δ = 73.5, both in D2O.
When the crude 12 was distilled (Kugelrohr, ∼125◦C/0.2 mm
Hg) the oil obtained was
a mixture of cis and trans 3,4-
dihydroxytetrahydrofuran (major) and 12 (minor) as revealed by 13C
and 1H NMR.38,39
Reaction of 13 with Na3AsO3
To a cooled (ice-water) solution of arsenic(III) oxide (198 mg, 1 mmol)
and sodium hydroxide (240 mg, 6 mmol) in water (0.5 mL) was added
through a septum 3,4-epoxy-1-butene 13 (168 µL, 2.4 mmol) and the
reaction mixture was stirred at 0◦C for 30 min and then at RT for 24 h.
TLC (MeOH/conc. NH3 4:1) of the clear viscous solution showed traces
of 13 (Rf 0.90), a spot at Rf 0.77, and AsO3− (Rf 0.58). Evaporation
and drying in vacuo over P2O5 for 4 days g3ave a sticky, glassy solid,
the 1H NMR (D2O, DSS) of which showed the signal of CH AsO2− at
2
2.15 ppm, of CH CH at 5.30 and ∼5.8 ppm, of trans CH CH at35.95
2
ppm, of cis CH CH at ∼6.5 ppm, and other signals in the region 1.8–
1
4.6 ppm. When trifluoroacetic acid was added to the H NMR sample
adjusting the pH to ∼2, the signal at 2.15 ppm disappeared but no signal
due to CH AsO3H2 was observed at ∼2.9 ppm. The new spectrum was
2
still too complicated to be analysed. When a sample of the glassy solid
was acidified to pH ∼7 with HCl, evaporated, dried and analysed by 1H
NMR (D2O, DSS) a weak signal at ∼2.2 ppm due to CH AsO2− and a
2
3
strong signal at ∼2.8 ppm due to CH AsO3H− were detected, meaning
2
that the monosodium salt of the arsonic acid was stable. Acidification
to pH 2 did not produce any free arsonic acid.
Preparation of 8 from Crude 10
To a solution of arsenic(III) oxide (1.832 g, 9.25 mmol As2O3) and sodium
hydroxide (2.220 g, 55.5 mmol) in water (4 mL) was added dropwise
during 2 h the crude epoxide 10 (2.836 g, purity 68%). The very viscous
solution was stirred at RT for 24 h. TLC (MeOH/conc. NH3 4:1) showed
the presence of tetritol (Rf 0.77), AsO3− (Rf 0.57), and the salt of 8
3
(Rf 0.30). To the cooled (cold water) solution was added dropwise 12
M hydrochloric acid (4.6 mL, 55.5 mmol) until a pH of 2 was reached
in order to neutralize the basic components of the mixture. Then, the
solution was diluted with water (80 mL), applied onto the strongly acidic
cation exchange resin Dowex AG 50W-X8 (H+form) (2.2 × 26 cm), and
eluted with water (1 L) collecting 250 mL fractions. The fourth fraction
gave a white solid (17 mg), which was m-chlorobenzoic acid (by IR). The

2758
M. A. Lala et al.
first three fractions were pooled, evaporated (rotary, 60◦C), and dried
in vacuo to give a glass (4.976 g). The glass dissolved in water (30 mL),
was applied portion-wise (5 × 6 mL) onto a column of the strong anion
exchange resin Dowex AG 1-X8 (acetate form) (3 × 32 cm) and eluted
with a) water (300 mL) and b) 1 M aqueous acetic acid (1 L) collecting
50 mL fractions. The water eluted As2O3 and tetritol (2.006 g), the
fractions 8-10 gave an unidentified compound (6 mg) running just above
the product 8 on TLC, the fractions 11–13 gave the product 8 (1.963 g)
as a glass, and the fractions 14–22 gave a solid (521 mg). Yields of 8: 46%
from 10 and 17% from 11. The glassy product 8 was insoluble in Et2O,
CHCl3, Me2CO, and MeOH. Trituration with warm methanol (∼5 mL)
transformed the glass into a white solid. M.p.: at ∼212◦C shrinks and at
218◦C decomposes. IR (KBr): 3376 vs, broad, 1638 m, 1400 mw, 1235 mw,
1070 s, 904 s, 768 s, 671 w. 1H NMR (D2O, DSS): δ = 2.82 and 2.91 (2 m,
4H, for the CH AsO3H2 of the two diastereoisomers); 3.58-3.79 (m, 6H,
2
for CH2OH and CH(OH)CH2OH of the two diastereoisomers); 4.15 and
4.22 (2 m, 2H, for CH(OH)CH2As of the two diastereoisomers). 13C NMR
(D2O, DSS), (ratio of intensities in parenthesis): δ = 41.0 and 41.2 (1:2.2)
for CH2AsO3H2, 64.9 and 65.1 (1:2.2) for CH2OH, 68.4 and 68.8 (3.1:1)
for CH(OH)CH2AsO3H2, 76.5 and 77.3 (2.7:1) for CH(OH)CH2OH.
For elemental analyses, to 100 mg (∼ 0.43 mmol) of impure 8 (con-
taining MeOH and AcOH) dissolved in water (0.5 mL) was added while
stirring a clear solution of recrystallized barium hydroxide octahydrate
(151 mg, 0.48 mmol) dissolved in methanol (10 mL). The precipitated
salt was centrifuged after 2 h, washed with boiling methanol (3 mL),
left at RT to cool, centrifuged and dried in vacuo over phosphorus pen-
toxide. The barium salt of 8 (138 mg, 87%) is a white non-hygroscopic
solid, which is moderately soluble in water. Calculated for C4H9O6AsBa
(365.37): C 13.15, H 2.48%; found: C 13.43 H 2.28%. IR (KBr): 3326
1
broad, vs, 1650 m, 1450 m, 1076 s, 1040 ms, 824 vs. H NMR (D2O,
DSS): δ = 2.15–2.34 (m, 2H, for the CH AsO2−); 3.55–3.85 (m, 3H, for
2
3
the CH (OH)CH(OH) ); 4.05-4.20 (m, 1H, for the CH(OH)CH2As). 13
C
2
NMR (D2O, DSS), (ratio of intensities in parenthesis): δ = 39.7 and 40.1
(1:3) for CH As; 65.0 and 65.4 (1:3) for CH2OH; 69.6 and 70.2 (3:1) for
2
CH(OH)CH2As; 76.9 and 77.5 (3:1) for CH(OH)CH2OH.
Preparation of 8 from 12
To a warm (water bath, 40◦C) solution of arsenic(III) oxide (540 mg, 2.73
mmol As2O3) and sodium hydroxide (654 mg, 16.35 mmol) in water (1.2
mL) was added dropwise during 1.5 h a light suspension of impure 12
(766 mg, 90% purity) in diethyl ether (1 mL) in a way allowing the di-
ethyl ether to evaporate and 12 to enter the aqueous phase. Stirring was

Diastereoisomeric 2,3,4-Trihydroxybutylarsonic Acids
2759
continued for 1 h at 40◦C and for further 24 h at RT. TLC (MeOH/conc.
NH3 4:1) revealed the presence of a large amount of tetritol (Rf 0.62),
AsO33− (Rf 0.43), and the salt of 8 (Rf 0.23). The cooled (cold water)
solution was acidified with 12 M hydrochloric acid (0.9 mL, 10.9 mmol)
to pH 2 and subjected to chromatography, as described above, using
the strong acidic and the strong basic resins. The product 8 was a glass
(441 mg), which on trituration with warm methanol (2 mL) gave a white
solid. Yields of 8: 31% from 12, 11% from 11. The IR, 1H and 13C NMR
spectra were the same as those obtained for the product prepared from
10.
For elemental analyses, the impure 8 (containing MeOH and AcOH)
was converted into its barium salt, as previously described. Yield 81%.
Calculated for C4H9O6AsBa (365.37): C 13.15, H 2.48%; found C 13.13,
H 2.48%. The IR, 1H and 13C NMR spectra of the barium salt of 8
produced from 12 were identical to those obtained for the product from
10.
REFERENCES
[1] V. M. Dembitsky and D. O. Levitsky, Prog. Lipid Res., 43, 403 (2004).
[2] T. Junk, G. C. Pappalardo, and K. J. Irgolic, Appl. Organomet. Chem., 4, 103 (1990).
[3] G. M. Tsivgoulis, D. N. Sotiropoulos, and P. V. Ioannou, Phosphorus, Sulfur, and
Silicon, 63, 329 (1991).
[4] S. V. Serves, G. M. Tsivgoulis, D. N. Sotiropoulos, and P. V. Ioannou, Phosphorus,
Sulfur, and Silicon, 71, 99 (1992).
[5] S. V. Serves, D. N. Sotiropoulos, P. V. Ioannou, and M. K. Jain, Phosphorus, Sulfur,
and Silicon, 81, 181 (1993).
[6] S. V. Serves, D. N. Sotiropoulos, and P. V. Ioannou, Phosphorus, Sulfur, and Silicon,
106, 75 (1995).
[7] A. Terzis and P. V. Ioannou, Chem. Phys. Lipids, 117, 53 (2002).
[8] N. L. Kordalis and P. V. Ioannou, Appl. Organomet. Chem., 14, 273 (2000).
[9] P. V. Ioannou, Chem. Phys. Lipids, 117, 7 (2002).
[10] O. Gortzi, P. Klepetsanis, S. G. Antimisiaris, and P. V. Ioannou, Chem. Phys. Lipids,
112, 21 (2001).
[11] D. G. Fatouros, O. Gortzi, P. Klepetsanis, S. G. Antimisiaris, M. C. A. Stuart, A.
Brisson, and P. V. Ioannou, Chem. Phys. Lipids, 109, 75 (2001).
[12] C. T. Supuran, S. V. Serves, and P. V. Ioannou, J. Inorg. Biochem., 62, 207 (1996).
[13] J. Rogers, B.-Z. Yu, S. V. Serves, G. M. Tsivgoulis, D. N. Sotiropoulos, P. V. Ioannou,
and M. K. Jain, Biochemistry, 35, 9375 (1996).
[14] O. Gortzi, E. Papadimitriou, C. G. Kontoyannis, S. G. Antimisiaris, and P. V. Ioannou,
Pharm. Res., 19, 79 (2002).
[15] O. Gortzi, S. G. Antimisiaris, P. Klepetsanis, E. Papadimitriou, and P. V. Ioannou,
Eur. J. Pharm. Sci., 18, 175 (2003).
[16] S. G. Antimisiaris, P. V. Ioannou, and P. M. Loiseau, J. Pharm. Pharmacol., 55, 647
(2003).
[17] D. G. Fatouros, P. V. Ioannou, and S. G. Antimisiaris, J. Nanosc. Nanotech., 6, 2618
(2006).

2760
M. A. Lala et al.
[18] S. L. Soignet, S. R. Frankel, D. Douer, M. S. Tallman, H. Hantarjian, E. Calleja, R.
M. Stone, M. Kalaycio, D. A. Scheinberg, P. Steinherz, E. L. Sievers, S. Coutre, S.
Dahlberg, R. Ellison, and R. P. Warrell, J. Clin. Oncol., 19, 3852 (2001).
[19] N. C. Munshi, G. Tricot, R. Desikan, A. Bardos, M. Zangari, A. Toor, C. Morris, E.
Anaissie, and B. Barlogie, Leukemia, 9, 1835 (2002).
[20] D. L. MacDonald, H. O. L. Fieser, and C. E. Ballou, J. Am. Chem. Soc., 78, 3720
(1956).
[21] J. W. Long and W. J. Ray Jr., Biochemistry, 12, 3932 (1973).
[22] P. V. Ioannou, Phosphorus, Sulfur, and Silicon, 177, 1 (2002).
[23] G. Meyer, Ber. Dtsch. Chem. Ges., 16, 1439 (1883).
[24] G. V. Chelintsev and V. K. Kushkov, J. Gen. Chem. USSR, 16, 1481 (1946); Chem.
Abstr., 41, 5441a (1947).
[25] C. K. Banks, J. F. Morgan, R. L. Clark, E. B. Hatlelid, F. H. Kahler, W. H. Paxton,
E. J. Cragoe, R. J. Andres, B. Elpern, R. F. Coles, J. Lawhead, and C. S. Hamilton,
J. Am. Chem. Soc., 69, 927 (1947).
[26] S. V. Serves, D. N. Sotiropoulos, P. V. Ioannou, and H. B. F. Dixon, Phosphorus,
Sulfur, and Silicon, 90, 103 (1994).
[27] G. M. Tsivgoulis, D. N. Sotiropoulos, and P. V. Ioannou, Phosphorus, Sulfur, and
Silicon, 57, 189 (1991).
[28] B. Damin, J. Garapon, and B. Sillion, Synthesis, 362, (1981).
[29] M. Wengert, A. M. Sanseverino, and M. C. S. de Mattos, J. Braz. Chem. Soc., 13, 700
(2002).
[30] L. S. Boguslavskaya, Russ. Chem. Rev., 41, 740 (1972).
[31] D. S. Mahadevappa and H. M. K. Naidu, Talanta, 20, 349 (1973).
[32] D. S. Mahadevappa and H. M. K. Naidu, Aust. J. Chem., 27, 1203 (1974).
[33] B. T. Golding, P. K. Slaich, and W. P. Watson, IARC Scientific Publication, No 70,
227 (1986).
[34] D. J. Claffey and J. A. Ruth, Tetrahedron Lett., 37, 7929 (1996).
[35] F. A. Long and J. G. Pritchard, J. Am. Chem. Soc., 78, 2663 (1956).
[36] J. G. Pritchard and F. A. Long, J. Am. Chem. Soc., 78, 2667 (1956).
[37] L. Smith and S. Skyle, Acta Chem. Scand., 4, 39 (1950).
[38] L. Que Jr. and G. R. Gray, Biochemistry, 13, 146 (1974).
[39] J. Skarewski and A. Gupta, Tetrahedron Asymmetry, 8, 1861 (1997).
[40] A. S. Rao, K. S. Pakniker, and J. G. Kirtane, Tetrahedron, 39, 2323 (1983).
[41] C. Jaime, R. M. Ortun˜o, and J. Font, J. Org. Chem., 53, 139 (1988).
[42] R. G. Kadesch, J. Am. Chem. Soc., 68, 41 (1946).
[43] H. B. F. Dixon, Adv. Inorg. Chem., 44, 191 (1997).
[44] C. Neagu and T. Hase, Tetrahedron Lett., 34, 1629 (1993).
[45] D. J. Claffey and J. A. Ruth, Tetrahedron Asymmetry, 8, 3715 (1997).
[46] A. -M. Lacoste, C. Dumora, B. R. S. Ali, E. Neuzil, and H. B. F. Dixon, J. Gen. Microb.,
138, 1283 (1992).
[47] S. V. Serves, D. N. Sotiropoulos, P. V. Ioannou, E. K. Mutenda, M. J. Sparkes, and
H. B. F. Dixon, Phosphorus, Sulfur, and Silicon, 101, 75 (1995).
[48] G. B. Payne, J. Org. Chem., 27, 3819 (1962).
[49] L. Ku¨rti and B. Czako´, Strategic Applications of Named Reactions in Organic Syn-
thesis (Elsevier, Amsterdam, 2005), pp. 336-337.
[50] W. F. Beech, J. Chem. Soc., 2483 (1951).
[51] P. W. Feit, Chem. Ber., 93, 116 (1960).
[52] A. I. Vogel, Practical Organic Chemistry (Longmans, London, 1964), 3rd edition,
Ch. IV, p. 824.
[53] T. H. Rider and A. J. Hill, J. Am. Chem. Soc., 52, 1521 (1930).