An efficient method for the refinement of 1,3-methyleneglycerol via bridged acetal exchange and the synthesis of a symmetrically branched glycerol trimer

Article abstract of DOI:10.1055/s-0031-1289809

Acid-catalyzed equilibrium of a mixture of 1,2- and 1,3-methyleneglycerol in 1,4-dioxane affords predominantly the 1,3-isomer via bridged acetal exchange. The minor 1,2-isomer is removed via sequential pivaloylation and tritylation to afford the desired 1,3-isomer in >99.5% purity. A symmetrically branched triglycerol is efficiently synthesized starting from the purified 1,3-isomer. Georg Thieme Verlag Stuttgart · New York.

Full text of DOI:10.1055/s-0031-1289809

PAPER
▌2365
paper
An Efficient Method for the Refinement of 1,3-Methyleneglycerol via Bridged
Acetal Exchange and the Synthesis of a Symmetrically Branched Glycerol
Trimer
Refinement of 1,3-Methyleneglycerol via Bridged Acetal Exchange
Hatsuhiko Hattori, Tsuyoshi Matsushita, Kohsuke Yoshitomi, Ayato Katagiri, Hisao Nemoto*
Department of Pharmaceutical Chemistry, Institute of Health Biosciences, Graduate School of The University of Tokushima, 1-78-1, Sho-machi,
Tokushima 770-8505, Japan
Fax +81(88)6337284; E-mail: nem@ph.tokushima-u.ac.jp
Received: 27.03.2012; Accepted after revision: 14.05.2012
HO
Abstract: Acid-catalyzed equilibrium of a mixture of 1,2- and 1,3-
methyleneglycerol in 1,4-dioxane affords predominantly the 1,3-
isomer via bridged acetal exchange. The minor 1,2-isomer is re-
moved via sequential pivaloylation and tritylation to afford the de-
sired 1,3-isomer in >99.5% purity. A symmetrically branched
triglycerol is efficiently synthesized starting from the purified 1,3-
isomer.
OH
OH
HO
O
HO
O
O
O
H
H
n
sublinear-
type
linear-type
OH
n
+ dendritic-type combined with linear and sublinear
Figure 2 Examples of other OPGLs
Key words: alcohols, acylation, isomerization, oligomerization,
glycerol
tected glycerol 1 was used as the key starting material for
all TP-BGLs.
Oligo- and polyglycerols possessing ether linkages (OP-
GLs) have been studied to improve various properties of
medicinal molecules.1 Among typical OPGLs,1 our devel-
oped series of symmetrically branched oligoglycerols
(BGLs), such as the trimer (BGL003)2 and heptamer
(BGL007),3 are more versatile for modification of a me-
dicinal molecule than other OPGLs. This is due to the fact
that BGLs are chemically pure (non-distributed mixtures),
and no asymmetric center is induced when the apex point
is connected to the target molecule (Figure 1). In contrast,
other OPGLs often exist as a distributed mixture of mo-
lecular weights with a number of uncontrolled asymmet-
ric centers (Figure 2).1
Protected glycerol 1 is not only an essential material for
the preparation of the BGL family of molecules, but is
also a versatile C3 building block in organic synthesis.5 As
such, many methods for the preparation of 1 have been re-
ported,2,6–8 which can be classified into two types
(Scheme 2). The first type results in the cyclic form 1A,
which is prepared from glycerol and an aldehyde or ke-
tone (R1COR2) (method A).6,7 The second leads to acyclic
form 1B, which is obtained from epichlorohydrin and an
alcohol (R3OH) (method B).2,8,9
Method A is more sustainably acceptable than method B
due to the following reasons. Industrial-scale production
of biofuels (methyl or ethyl esters of fatty acids) from veg-
etable triglycerides affords a tremendous amount of glyc-
erol as a surplus material.10 In contrast, epichlorohydrin is
prepared from petroleum (e.g., allylic chlorination of pro-
The strategy for the preparation of the terminus-protected
BGLs2–4 (TP-BGL) is illustrated in Scheme 1. 1,3-Pro-
terminus
terminus
OH OH OH OH
OH OH OH OH
OH
OH
OH
OH
O
O
O
O
O
O
O
O
apex
apex
OH
OH
BGL003
BGL007
Figure 1 Examples of our developed BGLs as OPGLs
SYNTHESIS 2012, 44, 2365–2373
Advanced online publication: 28.06.2012
0
0
3
9
-
7
8
8
1
1
4
3
7
-
2
1
0
X
DOI: 10.1055/s-0031-1289809; Art ID: SS-2012-F0315-OP
© Georg Thieme Verlag Stuttgart · New York

2366
H. Hattori et al.
PAPER
IMDA
TP-BGL006
TP-BGL012
TP-BGL014
IMDA
IMDA
OR OR OR OR
OR OR OR OR
OR OR OR OR
OR OR
ECH
ECH
O
O
O
O
O
O
OH
1
OH
TP-BGL003
O
O
ECH = epichlorohydrin
IMDA = N-protected iminodiacetic acid
TP- = terminus protected with an R group
OH
TP-BGL007
Scheme 1 Preparation of the BGL family of compounds
for the separation of the isomers would need to be devel-
oped. As shown in Scheme 3, however, undesired selec-
tivities using acetone (3:4 = >99:<1)13 and benzaldehyde
(5:6 = 50:50)7 were reported.
Based on the reported selectivities,7,13 we hypothesized
that the selectivity for 1A versus 2A would be very low if
both the R1 and R2 groups were large, due to 1,3-diaxial
repulsions13 in 1A (Figure 3).
O
ECH
O
O
R1
R2
OH
OH
R1
R2
TP-BGL003
HO
HO
method A
ECH
1A
TP-BGL007
O
O
O
Ph
H
cyclic-
type
HO
HO
O
H
R1
R1
R2
H
H
H
R2
OR3
OR3
O
Cl
O
H
ECH
HO
R3
O
O
TP-BGL003
HO
O
H
O
method B
O
2A
1A
1B
ECH
HO
Figure 3 Steric effects on the synthesis of 1,2- and 1,3-protected
glycerols
OBn
OBn
O-allyl
O-allyl
acyclic-
type
HO
Therefore, the selectivity for 1A might, in theory, be high
when both R1 and R2 are small (e.g. a hydrogen atom). Ini-
tially, however, we were disappointed to find that the mo-
lar ratios of both a literature reported mixture,14 and a
commercially available mixture of 7 and 8 were much
lower than expected.
Scheme 2 Methods for the preparation of 1,3-protected glycerol 1
pene followed by epoxidation). Furthermore, method A
comprises a dehydrating process while method B includes
a substitution reaction that essentially wastes a halogen at-
om.
Despite this, we focused our attention on protected glyc-
erol 7 because the methylene-bridged protecting group is
the smallest of all the versatile protecting groups for diols,
and demonstrates advantages of economy and sustainabil-
ity. Thus, we planned to separate isomer 7 from 8, on large
scale, using an acceptable procedure such as distillation or
recrystallization. These two attempts were, however, im-
mediately abandoned because the boiling point of 7
(193.8 °C) was very similar to that of 8 (192.5 °C),15 and
neither of their melting points were accessible (their melt-
ing points must be impractically low).
In addition, for our BGL project, method A is more syn-
thetically acceptable than method B because we can suc-
cessfully prepare TP-BGL007 bearing cyclic protecting
groups from the precursor TP-BGL003,3 while the reac-
tion of acyclic TP-BGL003 to give TP-BGL007 proved
unsuccessful. Additionally, the deprotecting conditions
for 1A are generally milder than those required for 1B.11
Accordingly, the use of method A is not only essential for
our BGL project, but is also advisable for synthetic studies
using the glyceryl skeleton as a C3 building block.5,12
Instead, we examined chemoselective acylation using piv-
aloyl chloride (PvCl); it is known that pivaloylation of a
primary hydroxy functionality is much faster than that of
a secondary one.16,17 Furthermore, the undesired piv-
To prepare 1A on large scale, the selectivity in favor of the
1,3-protected isomer versus the 1,2-protected isomer
should be as high as possible. Thus, a simple procedure
Synthesis 2012, 44, 2365–2373
© Georg Thieme Verlag Stuttgart · New York

PAPER
Refinement of 1,3-Methyleneglycerol via Bridged Acetal Exchange
2367
1,3-protection
O
1,2-protection
O
glycerol
HO
acetone
O
HO
H+, – H2O
O
4
3
<1:>99
glycerol
O
O
O
HO
benzaldehyde
H+, – H2O
HO
O
6
5
50:50
O
glycerol
O
O
formaldehyde
HO
HO
O
H+, – H2O
7
8
55:45 to 60:40
Scheme 3 Previously reported acid-catalyzed dehydration reactions of glycerol with an aldehyde or a ketone
aloylated esters can be easily recycled to the original alco- above mixture of 9 and 10 (9:91) to afford a mixture of 7
hols by alkaline hydrolysis (Table 1).17,18
and 8 in almost quantitative yield in a 9:91 ratio, along
with sodium pivalate in 98% isolated yield. Accordingly,
this method for the efficient separation of 8 from cGF to
prepare 7 in high purity proved successful, and included
the recycling of isomer 8 as well as the recovered sodium
pivalate.
Since we used commercially available glycerol formal
(cGF), which exists as a 56:44 mixture of 7 and 8, the
ideal results for our purpose based on the sum of 7 and 8
would be yields of 7 (56%), 8 (0%), 9 (0%) and 10 (44%)
using 0.44 equivalents of pivaloyl chloride. After several
attempts on small scale, we found satisfactory conditions Next, as shown in Scheme 4, we attempted the acid-cata-
that afforded 7 (59%, within error limits, or including the lyzed isomerization between 7 and 8 via intermediate 11
isomerization process from 8 into 7), 8 (7%), 9 (0%), and and related species, primarily because such bridged acetal
10 (33%), using pivaloyl chloride (0.55 equivalents) and exchange reactions in organic synthesis are often reported
pyridine (0.68 equivalents) in dichloromethane (0.15 M) in good to excellent yields.6,15,19 However, the reported
at 0 °C to room temperature over 18 hours. All the yields conditions were not applicable for our purposes due to an
1
were determined by calculation of the H NMR spectro- unusual problem for small molecules such as 7 and 8.
scopic peak areas based on mesitylene as an internal stan-
This problem was the low stoichiometry of the recovered
dard. Next, the large-scale pivaloylation of cGF (1.81
mixture of 7 and 8, as non-trivial amounts of unidentified
mol) was carried out under the same conditions to afford
by-products remained in the flask after distillation. Al-
though it was difficult to analyze all the components of the
distillation residue, oligomeric compounds such as 12, 13
a mixture of 7 and 8 (95:5) in 50% isolated yield and a
mixture of 9 and 10 (9:91) in 33% isolated yield after dis-
tillation. We carried out the alkaline hydrolysis of the
Table 1 Pivaloylation of cGF
O
O
O
O
Cl
O
7 + 8
7
+
8
+
+
O
O
O
O
(56:44)
pyridine
CH2Cl2
cGF
9
10
Method
7
8
9
10
Small scalea
59%
7%
0%
33%
Large scale (1.81 mol)b
Ideal case
50% (7:8 = 95:5), (7 = 48%, 8 = 2%)
56% 0%
33% (9:10 = 9:91), (9 = 3%, 10 = 30%)
0%
44%
a Determined by 1H NMR spectroscopy.
b Yield of isolated materials after distillation.
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2012, 44, 2365–2373

2368
H. Hattori et al.
PAPER
H
H+
O
O
H
H
H+
O
O
O
H
O
H
O
O
O
O
O
O
8
n
7
etc
+ H2O
– H2O
– H+
OH OH
OH
– H+
H
H
OH
OH
OH
OH
11
11
11
OH
R
OH
Scheme 4 A typical route for isomerization between 7 and 8
+ ROH
– ROH
H
R
O
RO
OR
O
and further oligomers were present in the inseparable mix-
ture (Scheme 5).
+ ROH
– ROH
OH OH
O
OH
Even under strict anhydrous conditions, oligomers such as
12 and 13 can theoretically be produced because both 7
and 8 leave behind protic residues. Therefore, methylene
bridges can be randomly transferred throughout to pro-
duce fully dehydrated compounds such as 13 with no free
hydroxy groups, and partially dehydrated compounds
such as 12 with two hydroxy groups. Furthermore, oligo-
merization is generally fast when the molecule is small.
OH
OH
Scheme 6 Mechanism for the formation of by-products and glycerol
in the presence of a large excess of protic molecules
OH
OH
HO
H
H
O
OH
O
O
+ glycerol
– glycerol
OH
O
or
Although addition of protic molecules such as water or al-
cohols (ROH) may prevent oligomerization reactions, this
leads to the formation of problematic by-products as illus-
trated in Scheme 6, and a non-protected triol (i.e., glycer-
ol).
O
OH
OH
OH
11
OH
OH
– glycerol
HO
+ glycerol
OH
– glycerol
+ glycerol
We eventually discovered that these problems could be
solved by the addition of glycerol as a source of protons.
As illustrated in Scheme 7, the alcoholysis of 11 using
glycerol was no longer a side reaction. Repeat bond for-
mation and cleavage between 7, 8, 11, 14 and glycerol can
occur smoothly to achieve equilibrium between 7 and 8.
At the same time, the formation of oligomers, especially
fully dehydrated molecules such as 13, is strongly inhibit-
ed since sufficient protons are available from the addition-
al glycerol.
OH
OH
O
11
O
14
Scheme 7 Strong inhibition of hydrolysis and oligomerization in the
presence of glycerol
listed in Table 2. Accordingly, we hypothesized that cGF
(56:44 ratio) may not be a thermodynamically controlled
mixture.14,20 We first examined the reaction of cGF (2.0
M in diethyl ether) at reflux for 18 hours in the presence
of 0.5 equivalents of glycerol and a catalytic amount of
sulfuric acid (Table 2, entry 1). However, the mass bal-
ance was still low (51%) in diethyl ether, probably be-
cause of the low solubility of glycerol in this solvent. At a
Based on the idea described above, we examined the acid-
catalyzed isomerization of a mixture of 7 and 8 (9:91 ra-
tio) in the presence of glycerol. As anticipated, the mate-
rial balance between the input and output of 7 and 8 was
greatly improved. Additionally, we made another unex-
pected and positive discovery: the ratio of 7 and 8, after
the acidic treatment, was often much more than 56:44 as
O
O
O
O
HO
O
– H+
further
oligomerizations
O
OH
O
OH
H+
+ 7
– 8
– 7
OH
11
OH
12
+ 8
intramolecular-
bridged dimer
O
O
O
O
O
HO
O
O
O
+ 8
O
O
O
OH
OH OH
O
13
intramolecular-
bridged dimer
OH
OH
Scheme 5 Typical examples of side reactions (self-oligomerization) due to the presence of protic residues in 7 and 8
Synthesis 2012, 44, 2365–2373 © Georg Thieme Verlag Stuttgart · New York

PAPER
Refinement of 1,3-Methyleneglycerol via Bridged Acetal Exchange
2369
low reaction temperature (17 °C), the ratio of 7:8 was un- periods were examined (3, 6, 18, and 48 h, Table 2, entries
changed, probably due to the reason that no deprotection 14, 15, 8 and 16). The results indicated that a reaction time
or re-protection processes occurred (Table 2, entry 2). between 18 and 48 hours was long enough to reach ther-
Other aromatic and ethereal solvents21 were examined un- modynamic equilibrium. Finally, reactions at various
der the same conditions (18 h, sulfuric acid, 60 °C) (Table temperatures were examined (not all results are shown in
2, entries 3–8). In each case, the molar ratio of 7 was in- Table 2). Below 50 °C, acid-catalyzed isomerization
creased, thus verifying our hypothesis. Accordingly, 1,4- starting from cGF was too slow (e.g., at 17 °C over 48 h,
dioxane was chosen as the most suitable solvent (Table 2, 7:8 = 59:41, Table 2, entry 17). Above 70 °C, the mass
entry 8) because it gave the highest mass balance of the balance was low and the ratio of 7 decreased (e.g., at 95
product.
°C over 18 h, 7:8 = 64:36, Table 2, entry 18). Thus, a tem-
perature around 60 °C was considered optimum to maxi-
mize the amount of the desired compound 7.
As anticipated, the reaction in the absence of glycerol
gave a decreased chemical yield (Table 2, entry 9), where-
as almost the same yields were obtained when 0.5, 1.0, Under the most suitable conditions (Table 2, entry 8),
and 2.0 equivalents of glycerol were used (Table 2, entries isomerization of cGF (7:8 = 56:44, 188 g, 1.81 mol) was
8, 10 and 11). Therefore, 0.5 equivalents of glycerol with carried out in the presence of glycerol (83.3 g, 0.91 mol)
respect to cGF were used in subsequent reactions.
at a higher concentration (5.0 M of cGF in 1,4-dioxane)
than the small-scale reaction in preparation for a future
large-scale reaction (Scheme 8). After distillation, a mix-
ture of 7 and 8 (70:30, 156 g, 1.50 mol) was afforded in
83% isolated yield. Glycerol was also recovered in 95%
yield by distillation. The resulting distillate (156 g of a
mixture of 7 and 8) was pivaloylated to afford 7 and 8
(97:3, 78.2 g, 0.75 mol, 41% yield based on starting cGF)
In entries 8, 12 and 13 (Table 2), the final molar ratios of
7:8 were the same within error limits, even for different
initial molar ratios of 7, that is, medium (56:44, Table 2,
entry 8), high (97:3, Table 2, entry 12) and low (30:70, Ta-
ble 2, entry 13). Therefore, the thermodynamic balance
between 7 and 8 at 60 °C in 1,4-dioxane was estimated to
be approximately 77:23.20 In addition, various reaction
Table 2 Acid-Catalyzed Bridged Acetal Exchange between 7 and 8a
Entry
Initial ratio
(7:8)
Additional glycerol Solventb
(vs 7 + 8)
Time
(h)
Temp
(°C)
Yieldc (%)
of 7 + 8
Ratio
(7:8)
1
2
56:44
56:44
56:44
56:44
56:44
56:44
56:44
56:44
56:44
56:44
56:44
97:3
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
–
Et2O
18
18
18
18
18
18
18
18
18
18
18
18
18
3
36
17
60
60
60
60
60
60
60
60
60
60
60
60
60
60
17
95
51
99
67
62
84
45
71
90
82
92
90
99
80
85
86
94
98
85
82:18
56:44
81:19
79:21
77:23
78:22
77:23
74:26
71:29
77:23
76:24
77:23
75:25
63:37
71:29
77:23
59:41
64:36
Et2O
3
neat
4
toluene
5
THF
6
MTBE
7
CPME
8
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
1,4-dioxane
9
10
11
12
13
14
15
16
17
18
1.0 equiv
2.0 equiv
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
0.5 equiv
30:70
56:44
56:44
56:44
56:44
56:44
6
48
18–48
18
a Reaction conditions: methyleneglycerols 7 and 8 (1.0 equiv), solvent (2.0 M based on methyleneglycerols), glycerol (0.5 equiv), H2SO4 (4
mol%) unless otherwise noted.
b MTBE = methyl tert-butyl ether, CPME = cyclopentyl methyl ether.
c Determined by 1H NMR spectroscopy using mesitylene as an internal standard.
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2012, 44, 2365–2373

2370
H. Hattori et al.
PAPER
H2SO4
glycerol
83.3 g
0.91 mol
glycerol
78.8 g
7 + 8
+
7 + 8
(70:30)
156 g
1.50 mol
83% yield
+
(56:44)
188 g
1,4-dioxane
60 °C, 18 h
0.86 mol
95% recovery
1.81 mol
O
starting cGF
O
NaO
recyclable
pyridine
CH2Cl2
Cl
NaOH
7 + 8
+
9 + 10
(9:91)
7 + 8
(97:3)
78.2 g
(9:91)
50.0 g
EtOH
93.2 g
0.75 mol
41% yield
0.50 mol
28% yield
0.48 mol
27% yield
7
pyridine
CH2Cl2
H2SO4
1,4-dioxane
60 °C, 18 h
(97% purity)
115.4 g
1.11 mol
61% yield
O
7 + 8
(97:3)
37.2 g
0.36 mol
20% yield
+
9 + 10
(5:95)
16.8 g
0.09 mol
5% yield
final output
Cl
all yields are indicated based on the starting cGF
Scheme 8 Preparation of compound 7 on large scale
after distillation. A mixture of 9 and 10 (9:91, 93.2 g, 0.50
mol, 28% isolated yield based on starting cGF) was also
obtained by distillation. After alkaline hydrolysis of the
mixture of 9 and 10 with sodium hydroxide (1.0 M) in eth-
anol, a mixture of 7 and 8 (9:91) was obtained in 27%
yield based on starting cGF (95% yield from the mixture
of 9 and 10). Repeated isomerization–pivaloylation af-
forded a mixture of 7 and 8 (97:3, 37.2 g, 0.36 mol, 72%
yield from the mixture of 9 and 10, 20% yield based on
starting cGF). A second portion of a mixture of 9 and 10
(5:95, 16.8 g, 89.3 mmol, 5% yield based on starting cGF)
was also obtained. Accordingly, cGF (1.81 mol) was con-
verted into 7 in 61% [= (0.75 + 0.36)/1.81] yield and 97%
purity.
1) TrCl (0.06 equiv)
2,4,6-collidine (0.12 equiv)
CH2Cl2 (1.0 M)
TrCl
OTr
7 + 8
(97:3)
O
N
2) NaHCO3 (aq)
O
O
O
TrOH, etc.
OTr
hexane layer
+
filtration
7
NaCl
NaHCO3
7
aqueous layer
distillation
(99.5–100% purity)
85–90% yield
Scheme 9 Final treatment to afford 7 in 99.5–100% purity
Although the purity of 7 obtained according to Scheme 8
was high enough for some purposes, a further procedure
to afford 7 in 99.5–100% purity is described. A solution
of 7 (97% purity) in dichloromethane was treated with tr-
ityl chloride22 (0.06 equiv) and 2,4,6-collidine (0.12
equiv) in dichloromethane (1.0 M). After several hours
(depending on the reaction scale), water containing 1%
sodium hydrogen carbonate was added and the dichloro-
methane was removed under reduced pressure. Since both
7 and 8 are soluble in water, tritylated compounds and
2,4,6-collidine were removed by extraction with hexane.23
After concentration of the aqueous layer and removal of
inorganic salts such as sodium chloride and sodium hy-
drogen carbonate by filtration, simple distillation then
gave the desired product 7 in 85–90% yield and in over
99.5% purity (Scheme 9).
mide and wet potassium hydroxide with vigorous
stirring2,3 to afford 15 in 90% yield. Subsequent hydroly-
sis of 15 with hydrochloric acid (1.0 M) in methanol gave
the pentaol 16, which was easily protected with 2,2-dime-
thoxypropane [Me2C(OMe)2] in the presence of a catalyt-
ic amount of Amberlyst 15 in N,N-dimethylformamide
to give 179 in 76% overall yield from 15. The final product
17 was purified both by silica gel column chromatography
and by recrystallization from diethyl ether–hexane (1:3) at
–10 °C (Scheme 10).
In conclusion, we have developed an efficient method for
the large-scale preparation of highly purified 1,3-methy-
leneglycerol 7. The main reagents (the pivaloyl moiety,
the sodium cation and additional glycerol) were recovered
in excellent yields. The synthesis of two different sym-
metrically branched trimers (TP-BGL003) was also suc-
cessful. At the same time, problematic side reactions
Finally, dendritic-generation growth of 7 was carried out
to afford the bis(methylene)-protected symmetrically
branched triglycerol 15. Purified 7 was reacted with epi-
chlorohydrin in the presence of tetrabutylammonium bro-
Synthesis 2012, 44, 2365–2373
© Georg Thieme Verlag Stuttgart · New York

PAPER
Refinement of 1,3-Methyleneglycerol via Bridged Acetal Exchange
2371
O
O
OH
O
O
O
O
OH
OH
O
O
O
O
O
O
a
c
b
HO
7
HO
HO
O
O
OH
16
BGL003(mtl)2
(15)
90% yield
BGL003(Atn)2
mtl = methylene
Atn = acetonide
(17)
76% yield from 15
Scheme 10 Synthesis of TP-BGL003 15 and 17. Reagents and conditions: (a) ECH, TBAB, KOH, neat, 60 °C, 48 h, 90%; (b) 1 M HCl, MeOH,
reflux, 24 h; then Amberlite IRA-96TM (c) Me2C(OMe)2, cat. Amberlyst 15, DMF, r.t., 12 h, 76% from 15.
(67.7 mL, 0.550 mol) dropwise over 45 min at 0 °C. After stirring
at r.t. for 6 h, the mixture was quenched with powdered NaHCO3
(ca. 237 mg, 2.24 mmol), and the resulting suspension filtered. The
filtrate was evaporated under reduced pressure to give a colorless
oil, which was distilled to furnish 7 in 50% yield with 97% purity
(65 °C/8 mmHg, 78.2 g, 0.751 mol), along with a mixture of 9 and
10 as a colorless oil (9:10 = 9:91, 93.2 g, 0.495 mol) (85 °C/8
mmHg) in 33% yield, which was then used for the next hydrolysis
procedure without further separation.
1H NMR (400 MHz, CDCl3): δ = 3.01 (d, J = 10.0 Hz, 1 H), 3.63
(dquin, J = 10.0, 3.0 Hz, 1 H), 3.88 (dd, J = 11.0, 3.0 Hz, 2 H), 3.93
(dd, J = 11.0, 3.0 Hz, 2 H), 4.77 (d, J = 6.0 Hz, 1 H), 4.94 (d, J = 6.0
Hz, 1 H).
involving acid-catalyzed exchange of acetal bridges in
small molecules were addressed, and a novel solution (ad-
dition of the fully deprotected original alcohol if readily
available) was demonstrated.
All reactions were carried out under an argon atmosphere unless
otherwise described. Toluene, Et2O, 1,4-dioxane, DMF, CH2Cl2
and 2,2-dimethoxypropane were distilled using standard methods
prior to use. Glycerol formal (cGF) was purchased from Tokyo
Chemical Industry Co., Ltd., and was found to contain 56% of 1,3-
methyleneglycerol (4-hydroxy-1,3-dioxane) (7) and 44% of 1,2-
methyleneglycerol (4-hydroxymethyl-1,3-dioxolane) (8) according
1
to H NMR spectroscopic analysis. Column chromatography was
13C NMR (75 MHz, CDCl3): δ = 64.2 (CHOH), 71.8 (2 × CH2), 94.1
(OCH2O).
ESI-HRMS: m/z [M + Na]+ calcd for C4H8O3Na: 127.0366; found:
127.0338.
performed using Silica gel 60N purchased from Kanto Chemical
Co., Inc. (63–210 m). IR spectra were obtained using a Jasco FT-
IR 6200 spectrophotometer.1H and 13C NMR spectra were recorded
using JEOL-EX400 or Bruker AV400n instruments and were refer-
enced to tetramethylsilane. High-resolution mass spectra were ob-
tained on a Waters LCT Premier spectrometer.
Hydrolysis of a Mixture of Compounds 9 and 10
To a soln of 9 and 10 (9:10 = 9:91, 29.1 g, 155 mmol) in EtOH (50
mL) was added an aq soln of NaOH (1 M, 233 mL, 233 mmol) at
r.t. After stirring at r.t. for 18 h, the mixture was neutralized by the
addition of HCl (1 M), and then evaporated under reduced pressure.
The resulting suspension containing NaCl and sodium pivalate was
filtered through a pad of Celite® which was then rinsed with
CH2Cl2. The combined organic and aqueous filtrates were evaporat-
ed in vacuo to remove CH2Cl2 and H2O to give a colorless oil, which
was distilled (65 °C/8 mmHg) to furnish a mixture of 7 and 8 (7:8 =
9:91, 15.3 g, 147 mmol).
Isomerization of 1,2-Methyleneglycerol; Typical Procedure
To a soln of methyleneglycerols 7 and 8 (520 mg, 5.0 mmol) and
glycerol (200 mg, 2.5 mmol) in 1,4-dioxane (2.5 mL) was added
concd H2SO4 (22.0 mg, 0.224 mmol) at r.t. After stirring at 60 °C
for 18 h, the mixture was cooled to r.t. and then quenched with pow-
dered Na2CO3 (ca. 237 mg, 2.24 mmol) with vigorous stirring. The
resulting suspension was filtered through a pad of Celite® which
was then rinsed with Et2O. The filtrate was evaporated under re-
duced pressure to give a crude material, which was analyzed by 1H
NMR spectroscopy after the addition of mesitylene (200 mg, 1.67
mmol) as an internal standard.
Tritylation of a Mixture of 7 and 8
To a stirred soln of a mixture of 7 and 8 (ratio = 97:3, 10.41 g, 0.10
mol) in CH2Cl2 (100 mL) in the presence of 2,4,6-collidine (1.09 g,
9.0 mmol) was added slowly a CH2Cl2 (10 mL) soln of TrCl (1.67
g, 6.0 mmol) in a dropwise manner at r.t., and then the resulting
mixture was stirred for ca. 7 h. The mixture was quenched with an
aq soln of NaHCO3 (1 wt%, 500 g, 60.0 mmol). The separated or-
ganic layer was concentrated in vacuo to remove CH2Cl2. The resi-
due and the separated aq layer were mixed together, and extracted
with hexane (2 × 50 mL) to remove lipophilic compounds such as
2,4,6-collidine and tritylated products. The aq layer was concentrat-
ed in vacuo in the presence of EtOH (as an azeotrope). [If the inor-
ganic salts (NaCl and NaHCO3) precipitated, filtration was carried
out to remove the salts in order to prevent bumping.] The residue
was dissolved in CH2Cl2 (200 mL), dried over anhyd MgSO4, and
concentrated in vacuo. The residue was distilled (66 °C/8 mmHg) to
afford 7 in 99.5–100% purity (8.85–9.37 g, 85.0–90.0 mmol, 85–
90% yield).
Large-Scale Isomerization of Methyleneglycerols; Typical Pro-
cedure
To a soln of cGF (188 g, 1.81 mol) and glycerol (83 g, 0.90 mol) in
1,4-dioxane (360 mL) was added concd H2SO4 (66.0 mg, 0.672
mmol) at r.t. After stirring at 60 °C for 18 h, the mixture was cooled
to r.t. and then quenched with powdered Na2CO3 (1.42 g, 13.44
mmol). The resulting suspension was filtered through a pad of
Celite® which was then rinsed with Et2O. The filtrate was evaporat-
ed under reduced pressure to give a crude residue, which was dis-
tilled (65 °C/8 mmHg) to furnish a mixture of 7 and 8 (70:30, 156
g, 1.50 mol, 83% yield). Further distillation (88 °C/0.1 mmHg) af-
forded glycerol (79.2 g, 0.86 mol, 96% recovered yield) as a color-
less oil.
1,3-Methyleneglycerol (7) (in 97% Purity)
To a soln of methyleneglycerols (7:8 = 70:30, 156 g, 1.50 mol) and
pyridine (48.0 mL, 0.600 mol) in CH2Cl2 (500 mL) was added PvCl
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2012, 44, 2365–2373

2372
H. Hattori et al.
PAPER
1,3-Bis[(1,3-dioxan-5-yl)oxy]propan-2-ol [15, BGL003(mtl)2]
To a suspension of TBAB (0.457 g, 1.42 mmol), 7 (2.21 g, 17.7
mmol) and KOH (0.994 g, 17.7 mmol) in H2O (0.5 mL) was added
epichlorohydrin (0.354 mL, 7.09 mmol) dropwise at r.t. with vigor-
ous stirring. After stirring at 60 °C for 48 h, the mixture was neu-
tralized with an aq soln of HCl (1.0 M), and evaporated under
reduced pressure to give a suspended material. The suspension con-
taining KCl was filtered through a pad of Celite® which was then
rinsed with 1,4-dioxane. The filtrate was evaporated in vacuo to af-
ford a crude oil, which was purified by silica gel column chroma-
tography (CH2Cl2–acetone, 3:1) to furnish 15 (1.69 g, 6.40 mol,
90% yield) as a colorless oil.
1H NMR (400 MHz, CDCl3): δ = 2.83 (d, J = 4.0 Hz, 1 H), 3.42–
3.49 (m, 2 H), 3.57 (dd, J = 9.0, 6.0 Hz, 4 H), 3.62 (dd, J = 9.0, 4.4
Hz, 4 H), 3.74 (dd, J = 12.0, 6.0 Hz, 4 H), 3.94 (ddd, J = 6.0, 4.4,
4.0 Hz, 1 H), 4.03 (dd, J = 12.0, 4.0 Hz, 4 H), 4.75 (d, J = 6.0 Hz, 2
H), 4.84 (d, J = 6.0 Hz, 2 H).
References
(1) For applications and preparations of oligoglycerols, see:
(a) Jayaraman, M.; Fréchet, J. M. J. J. Am. Chem. Soc. 1998,
120, 12996. (b) Grayson, S. M.; Fréchet, J. M. J. Chem. Rev.
2001, 101, 3819. (c) Haag, R.; Sunder, A.; Stumbé, J.-F.
J. Am. Chem. Soc. 2000, 122, 2954. (d) Calderón, M.;
Quadir, M. A.; Sharma, S. K.; Haag, R. Adv. Mater. 2010,
22, 190. (e) Calderón, M.; Quadir, M. A.; Strumia, M.;
Haag, R. Biochimie 2010, 92, 1242. (f) Allcock, H. R.;
Ravikiran, R.; O’Connor, S. J. M. Macromolecules 1997, 30,
3184. (g) Cassel, S.; Debaig, C.; Benvegnu, T.; Chaimbault,
P.; Lafosse, M.; Plusequellec, D.; Rollin, P. Eur. J. Org.
Chem. 2001, 875.
(2) Nemoto, H.; Wilson, J. G.; Nakamura, H.; Yamamoto, Y.
J. Org. Chem. 1992, 57, 435.
(3) Nemoto, H.; Kamiya, M.; Minami, Y.; Araki, T.;
Kawamura, T. Synlett 2007, 2091.
13C NMR (100 MHz, CDCl3): δ = 69.1 (4 × CH2O), 69.6 (CHOH),
70.1 (2 × CH2O), 70.6 (2 × CH), 94.1 (2 × OCH2O).
ESI-HRMS: m/z [M + Na]+ calcd for C11H20O7Na: 287.1106; found:
287.1096.
(4) (a) BGL002 and BGL006: Nemoto, H.; Araki, T.; Kamiya,
M.; Kawamura, T.; Hino, T. Eur. J. Org. Chem. 2007, 3003.
(b) BGL012: Ishihara, A.; Yamauchi, M.; Kusano, H.;
Mimura, Y.; Nakakura, M.; Kamiya, M.; Katagiri, A.;
Kawano, M.; Nemoto, H.; Suzawa, T.; Yamasaki, M. Int. J.
Pharm. 2010, 391, 237. (c) BGL014: Nemoto, H.; Ishihara,
A.; Araki, T.; Katagiri, A.; Kamiya, M.; Matsushita, T.;
Hattori, H.; Mimura, Y.; Tomoda, Y.; Yamasaki, M. Bioorg.
Med. Chem. Lett. 2011, 21, 4724.
(5) (a) Ghosh, A. K.; Gemma, S.; Baldridge, A.; Wang, Y.-F.;
Kovalevsky, A. Y.; Koh, Y.; Weber, I. T.; Mitsuya, H.
J. Med. Chem. 2008, 51, 6021. (b) Ichikawa, T.; Kitazaki,
T.; Matsushita, Y.; Yamada, M.; Hayashi, R.; Yamaguchi,
M.; Kiyota, Y.; Okonogi, K.; Itoh, K. Chem. Pharm. Bull.
2001, 49, 1102. (c) Gras, J.-L.; Bonfanti, J.-F. Synlett 2000,
248. (d) Hoffmann, R. H.; Schäfer, F.; Haeberlin, E.; Rohde,
T.; Körber, K. Synthesis 2000, 2060. (e) García, N.;
Compañ, V.; Díaz-Calleja, R.; Guzmán, J.; Riande, E.
Polymer 2000, 41, 6603. (f) Beugelmans, R.; Lechevallier,
A.; Gharbaoui, T.; Frinault, T.; Benhida, R. Chem. Lett.
1995, 24, 243.
1,3-Bis[(2,2-dimethyl-1,3-dioxan-5-yl)oxy]propan-2-ol [17,
BGL003(Atn)2]
A soln of 15 (0.528 g, 2.00 mmol) and HCl (1 M in MeOH; 2 mL)
was stirred at reflux for 24 h. The mixture was neutralized by the ad-
dition of Amberlite IRA-96™ (0.913 mmol/g; 4.38 g, 4 mmol). The
resin was removed by filtration and the filtrate was evaporated un-
der reduced pressure to give crude 16, which was used without fur-
ther purification.
To a soln of 16 and 2,2-dimethoxypropane (0.738 mL, 6.00 mmol)
in DMF (2 mL) was added Amberlyst 15® (39.5 mg, 0.200 mmol,
loading 5.06 mmol/g) portionwise. After stirring at r.t. for 24 h, the
mixture was filtered, and evaporated under reduced pressure to give
a brownish oil, which was purified by silica gel column chromatog-
raphy (CH2Cl2–acetone, 2:1) to furnish 17 (0.487 g, 1.52 mmol,
76% yield) as a colorless gummy solid. The spectral data of 17 were
in complete agreement with those of the same compound reported
in our previous work.3,9
(6) Acetonide: Forbes, D. D.; Ene, D. G.; Doyle, M. P. Synthesis
1998, 879.
IR (neat): 3477, 2975, 2917, 2860, 1478, 1460, 1447, 1183, 1155,
1078, 1040, 1017, 974, 942, 912, 802 cm–1.
1H NMR (400 MHz, CDCl3): δ = 1.41 (s, 6 H), 1.43 (s, 6 H), 2.68
(br s, 1 H), 3.42–3.49 (m, 2 H), 3.53 (dd, J = 9.6, 6.0 Hz, 2 H), 3.60
(dd, J = 9.6, 4.4 Hz, 2 H), 3.77 (dd, J = 12.6, 4.4 Hz, 4 H), 3.93 (br
s, 1 H), 3.98 (dd, J = 12.6, 6.0 Hz, 4 H).
13C NMR (100 MHz, CDCl3): δ = 23.5 (2 × CH3), 23.6 (2 × CH3),
62.5 (4 × CH2O), 69.7 (CHOH), 69.8 (2 × CH2O), 71.1 (2 × CH),
98.3 (2 × C).
(7) Benzylidene: (a) Crich, D.; Beckwith, A. L. J.; Chen, C.;
Yao, Q.; Davison, I. G. E.; Longmore, R. W.; De Parrodi, C.
A.; Quintero-Cortes, L.; Sandoval-Ramirez, J. J. Am. Chem.
Soc. 1995, 117, 8757. (b) Carlsen, P. H. J.; Sørbye, K.;
Ulven, T.; Aasbø, K. Acta Chem. Scand. 1996, 50, 185.
(8) Both 1,3-diallylglycerol [trade name: glycerol α,α′-diallyl
ether, trade code: D2146 ($164/mol)] and 1,3-dibenzyl-
glycerol [trade name: 1,3-bis(benzyloxy)-2-propanol, trade
code: B2110 ($2277/mol)] are commercially available.
(9) Nemoto, H.; Kamiya, M.; Nakamoto, A.; Katagiri, A.;
Yoshitomi, K.; Kawamura, T.; Hattori, H. Chem. Lett. 2010,
39, 856.
ESI-HRMS: m/z [M + Na]+ calcd for C15H28O7Na: 343.1733; found:
343.1740.
(10) For the uses of glycerol, see: (a) Zhou, C.-H.; Beltramini, J.
N.; Fan, Y.-X.; Lu, G. Q. Chem. Soc. Rev. 2008, 37, 527.
(b) Mota, C. J. A.; Da Silva, C. S. A.; Rosenbach, N.; Costa,
J. Jr.; Da Silva, F. Energy Fuels 2010, 24, 2733. (c) Da Silva,
C. X. A.; Gonçalves, V. L. C.; Mota, C. J. A. Green Chem.
2009, 11, 38. (d) Behr, A.; Eilting, J.; Irawadi, K.;
Leschinski, J.; Lindner, F. Green Chem. 2008, 10, 13.
(e) Ott, L.; Bicker, M.; Vogel, H. Green Chem. 2006, 8, 214.
(f) Ciriminna, R.; Pagliaro, M. Adv. Synth. Catal. 2003, 345,
383.
Acknowledgment
We thank the Japan Science and Technology Agency (JST) for par-
tial support for this project (Research for Adaptable and Seamless
Technology Transfer Program through Target-driven R&D, JST in
2009–2011).
Supporting Information for this article is available online at
http://www.thieme-connect.com/ejournals/toc/synthesis.SnuIomprifgtooatrinSugpIoi
nnfomartit
(11) Wuts, P. G.; Greene, T. W. Greene’s Protective Groups in
Organic Synthesis; John Wiley & Sons: New York, 2007.
(12) Method B produces 1,3-protected glycerol as the sole
product. In contrast, method A usually affords a mixture of
Synthesis 2012, 44, 2365–2373
© Georg Thieme Verlag Stuttgart · New York

PAPER
Refinement of 1,3-Methyleneglycerol via Bridged Acetal Exchange
2373
1,2- and 1,3-protected glycerol. This is the only advantage of
method B versus A.
(20) The hypothesis shown in Figure 3 becomes more credible as
a consequence of the results reported herein, because the
results of entries 8, 12 and 13 in Table 1 indicate objectively
the balanced point of thermodynamic equilibrium. The
conditions for the industrial preparation of cGF must
undoubtedly be different from our reported conditions,
although the industrial preparation was not described in
detail.
(21) Undesired compound 8 was obtained as the major isomer
when halogenated solvents such as CH2Cl2 or DCE were
chosen for the acid-catalyzed re-bridge reaction of cGF.
Additionally, wet ethereal solvents and/or wet glycerol
afforded less 7 and more 8 than the result obtained in entry
8. Therefore, the thermodynamic ratio (7:8 = 77:23) was
reached as long as anhydrous conditions were carefully
maintained.
(13) Showler, A. J.; Darley, P. A. Chem. Rev. 1967, 67, 427.
(14) During the course of our studies, the following paper was
published: Ruiz, V. R.; Velty, A.; Santos, L. L.; Leyva-
Pérez, A.; Sabater, M. J.; Iborra, S.; Corma, A. J. Catal.
2010, 271, 351. Although similar selectivity (7:8 = >75:<25)
was reported under some reaction conditions, the chemical
yield was always low in such cases. The procedure for the
separation of 7 and 8 was neither mentioned nor
demonstrated.
(15) Mukai, C.; Miyakawa, M.; Hanaoka, M. J. Chem. Soc.,
Perkin Trans. 1 1997, 913.
(16) Ac2O: Our initial attempt using Ac2O was unsuccessful as
very poor acetylation selectivity between 7 and 8 was
observed.
(17) TsCl: Although the use of TsCl for selective tosylation of 8
was reported, the recovered yield of 7 was very low and a
serious problem was that tosylated 7 and 8 could not be
recycled, see: Gras, J.-L.; Nouguier, R.; Mchich, M.
Tetrahedron Lett. 1987, 28, 6601.
(18) TrCl: The reaction of cGF with trityl chloride was
abandoned because deprotecting conditions to recycle the
reagent must be carefully controlled for tritylated 7 and 8 in
the presence of acetal protecting groups. Alternatively,
exhaustive hydrolysis to afford glycerol, formaldehyde and
trityl alcohol must be performed.
(19) (a) Coe, J. W.; Roush, W. R. J. Org. Chem. 1989, 54, 915.
(b) Krohn, K.; Börner, G. J. Org. Chem. 1994, 56, 6063.
(c) Zhu, J.; Ma, D. Angew. Chem. Int. Ed. 2003, 42, 5348.
(d) Prasad, K. R.; Chandrakumar, A. J. Org. Chem. 2007,
72, 6312.
(22) Treatment of compound 7 (of 97% purity) with TrCl
afforded 7 in much higher purity than similar treatment with
PvCl. Although we pointed out that tritylation of
methyleneglycerol was problematic if recycling results of
25–44% of 8 were considered,18 using TrCl to remove the
very small amount of 8 (Scheme 9) was acceptable for our
purposes.
(23) Extraction is often more economical and sustainable than
distillation since time and energy (electricity for running the
vacuum pump and heating equipment) can be saved.
Therefore, the separation procedure by extraction with
hexane was also examined for removing pivaloylated
compounds from 7 and 8, which proved to be as successful
as the tritylation procedure.
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2012, 44, 2365–2373