Synthesis and structure of 2-aryl-5,5-disubstituted-1,3-dioxanes and conversion into chiral (1,1,1-trishydroxymethyl) methane derivatives

Article abstract of DOI:10.1016/S0040-4039(02)00176-4

Pentaerythritol, (1,1,1-trishydroxymethyl)methyl methane and (1,1,1-trishydroxymethyl)nitromethane are converted into 2-aryl-5,5-bis(hydroxymethyl), 2-aryl-5-hydroxymethyl-5-methyl- or 2-aryl-5-hydroxymethyl-5-nitro-1,3-dioxanes and a range of derivatives. X-Ray and NMR analysis establishes that the latter is obtained as a single diastereomer whose structure is unambiguously determined. These materials can be elaborated to chiral derivatives of the starting (1,1,1-trishydroxymethyl) methanes.

Full text of DOI:10.1016/S0040-4039(02)00176-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 2091–2094
Synthesis and structure of 2-aryl-5,5-disubstituted-1,3-dioxanes
and conversion into chiral (1,1,1-trishydroxymethyl)
methane derivatives
John M. Gardiner,a,* Paul Mather,a Ramy Morjan,a Robin G. Pritchard,a John E. Warren,a
Malcolm L. Cooper,a Abd El-Rahman S. Ferwanahb and Omar S. Abu-Tiemb
aDepartment of Chemistry, UMIST, Manchester M60 1QD, UK
bDepartment of Chemistry, Al-Azhar University of Gaza, PO Box 1277, Gaza, Palestine
Received 6 November 2001; revised 4 January 2002; accepted 24 January 2002
Abstract—Pentaerythritol, (1,1,1-trishydroxymethyl)methyl methane and (1,1,1-trishydroxymethyl)nitromethane are converted
into 2-aryl-5,5-bis(hydroxymethyl), 2-aryl-5-hydroxymethyl-5-methyl- or 2-aryl-5-hydroxymethyl-5-nitro-1,3-dioxanes and a range
of derivatives. X-Ray and NMR analysis establishes that the latter is obtained as a single diastereomer whose structure is
unambiguously determined. These materials can be elaborated to chiral derivatives of the starting (1,1,1-trishydroxymethyl)
methanes. © 2002 Elsevier Science Ltd. All rights reserved.
Pentaerythritol (1), (1,1,1-trishydroxymethyl)methyl
methane (2) and (1,1,1-trishydroxymethyl)nitromethane
(3) are readily available and inexpensive materials.
They have numerous potential uses for synthesis.1 We
have been interested in elaborating these materials to
fully differentiated multifunctional intermediates, spe-
cifically of type 6 (i.e. with at least two of R1–R3 being
protected with orthogonal protecting groups), with at
least three differentiated functionalities for further elab-
oration. We report here the conversion of 2 and 3 into
such chiral derivatives and provide structural proof of
the diastereochemical outcomes of intermediate dioxane
formation with 3.
one group, and thus work proceeded by preparation of
the known3 orthoformate 7 (using triethyl orthofor-
mate). The one remaining free hydroxyl was derivatized
as its acetate or as a t-butyldiphenyl silyl ether, and in
each case the orthoformate was removed using water to
give 4 or 5, respectively.4 In both cases, the main issue
was the incomplete removal of the orthoformate pro-
tecting group. We ascertained that this could be largely
controlled in the case of acetate 4. The monoacetyl
pentaerythritol (4) could be obtained in about 80%
yield on prolonged warming during hydrolysis. How-
ever, previously when we conducted the aqueous cleav-
age reaction without heating, the unknown formyl
derivative 8 was obtained in 90% yield.
All approaches used the potential of benzylidene or
p-methoxybenzylidene acetals to select two hydroxy-
methyl groups, with subsequent reductive acetal half-
opening to differentiate the two acetal oxygens.
Pentaerythritol has four equivalent groups and so an
initial further differentiation is required. Although it is
possible to obtain monoacetals by direct reaction of
pentaerythritol, and subsequently to differentiate the
two remaining hydroxyls, product separations due to
the over-protection potential at each step are generally
required.2 The obvious strategy is initially to derivatize
In the case of the silylated derivative 5, however, even
on prolonged heating, a mixture of triol 5 and the novel
formyl diol 9 was always obtained. Purified 4 and 5
were converted to the corresponding benzylidene
acetals 10 and 12, respectively, but yields were variable,
around 30–40% after purification. In the case of the
reaction forming 10, this was in part due to concomi-
* Corresponding author. Tel.: +44 (0)161 200 4530; fax: +001 (510)
217 2371; e-mail: john.gardiner@umist.ac.uk
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00176-4

2092
J. M. Gardiner et al. / Tetrahedron Letters 43 (2002) 2091–2094
tant formation of the spirocyclic bis-acetal 11 (requiring
in situ deacylation). This presumably exists as the sym-
metrical doubly equatorial spirocyclic structure shown
(Schemes 1 and 2).5
Lack of reported data for, or proof of, related
stereostructures, encouraged us to establish unambigu-
ously this stereostructure. Establishing this in one case
would, by the NMR analogy between them, also sup-
port assignment of the other acetal. Proof of the iso-
meric structure was provided by X-ray structure
analysis of the t-butyldimethyl silyl ether derivative 20
[R=TBDMS] (Fig. 1).
However, the modest yields and variable reproducabil-
ity of efficient benzylidene acetal formation terminated
pursual of the pentaerythritol route. (1,1,1-Trishydroxy-
methyl)methyl methane (2) and (1,1,1-trishydroxy-
methyl)nitromethane (3) were chosen for evaluation
since one group is already differentiated thus circum-
venting the initial part of the pentaerythritol elabora-
tion. The same basic strategy was applied, with direct
conversion of these triols to the cyclic acetals, using, in
this case, either benzaldehyde or 4-methoxy-
benzaldehyde.
This proved that the single diastereomer formed in this
case is as shown for 16, with nitro and aryl rings cis,
and the nitro axial. By NMR comparison this also
confirms the stereostructure as shown for 15. Although
this molecule contains a plane of symmetry, the crystal
structure shows this does not pack with this symmetry.
The aryl ring is twisted relative to the O1–O3–C4–C6
plane. The preferences of 5-substituents on 1,3-dioxanes
have been the subject of various conformational analyt-
ical studies over the years.11 The nitro group favours an
axial orientation in the analogue of 15 lacking the
5-hydroxymethyl. The rationale for this preference is
the electrostatic interaction of the endocyclic oxygen
lone pairs and the nitro nitrogen.12
The reaction of (1,1,1-trishydroxymethyl)methyl meth-
ane with benzaldehyde generated a diastereoisomeric
mixture (typically 3:1 or greater). Assignment can be
made by comparison with data and analysis reported
by Stoddart6a and also by others.6b,c The major isomer
is 14a, with phenyl and hydroxymethyl groups cis, and
reported NMR data for each isomer corresponded with
literature. We observed the same outcome for the syn-
thesis of 13b/14b.
All the acetals were elaborated by protection of the
remaining free hydroxyl with various protecting groups,
to allow, ultimately, evaluation of choices of residual
protecting groups in the target differentiated systems
(and to enhance the options for orthogonality of pro-
tecting groups). These are shown in Table 1 for the
synthesis of 17–20 with five different protecting groups
introduced overall. Additionally, 2-carbon and 3-car-
bon ether extended analogues of 17 have been prepared
(Table 1), which should facilitate synthesis of further
analogues of 21–23 with extended and alternatively
functionalized groups.
Reaction of (1,1,1-trishydroxymethyl)nitromethane (3)
under the same conditions generated 157 or 16, but in
both cases as a single diastereomer in high yields.8
There are previous literature reports of these acetals,
but no data appears to have been reported,9 so assign-
ment by comparison was unavailable. The 1H reso-
nances for the three methylene groups in 15 and 16 are
essentially identical, indicating they have the same
diastereomeric structure. Interestingly, the t-butyl ace-
tal analogue has also been previously reported, with
data in that case showing a diastereomeric mixture of
acetals (Scheme 3).10
All three acetals 17, 19 and 20 [R=TBDMS] (reacted
with DIBAL-H) give cleanly high yields of the corre-
sponding chiral tetrafunctional methanes 21–23
(Scheme 4). This thus provides a convenient and scal-
able access to two families of chiral systems of this sort
(bearing a methyl or a nitro).
We have prepared a range of diastereomeric derivatives
(e.g. 24–26) to attempt either crystallization or chro-
matographic separation of the ultimate enantiomers,
but to date none has proven cleanly separable. We also
prepared a glucoside derivative of 22. We do, however,
Scheme 1. (i) AcCl, Pyr; (ii) H2O, heat; (iii) TBDPSCl, ImH.
Scheme 2. (i) PhCHO, CSA, MeCN.
Scheme 3.

J. M. Gardiner et al. / Tetrahedron Letters 43 (2002) 2091–2094
2093
Figure 1. X-Ray structure of 20 [R=TBDMS]. Lower representation shows view from C2 end in the O1,O3,C4,C6 plane.
Table 1. 17 and 19 refer to diastereomeric mixtures, whilst
18 and 20 are single diastereomers as illustration
X
Ar
Ph
R
Yield (%)
17
Me
SEM, MEM, TBDMS, allyl 65–75
TBDMSO(CH2)n [n=2, 3]
TBDMS
30–65a
70
18
19
20
NO2 Ph
Me
NO2 PMP
PMP
MEM, TBDMS, TBDPS
MEM, TBDMS
60–70
60–70
a For n=3, 60–65%; for n=2 yield is over four steps.13
In summary, several new derivatives of pentaerythritol
are reported, and selective routes to provide novel fully
differentiated, and thus chiral, derivatives of (1,1,1-
trishydroxymethyl)methyl methane (2) and (1,1,1-
trishydroxymethyl)nitromethane (3) are described. The
latter clearly offers scope for numerous elaborations
and evaluation of enantioseparation methods is
underway.
Acknowledgements
Scheme 4.
Arab Student Aid International and the British-Arab
Charitable Foundation are thanked for support to
R.M. The British Council are also thanked for support-
ing a prior exchange visit (R.M.). NMR, HPLC and IR
characterization used instrumentation funded by
EPSRC grants GR/L52246 (NMR), GR/M30135 (IR)
and GR/L84391 (HPLC). Julie Simkins is thanked for
some early preliminary work.
now have alternative access to a key homochiral nitro
system analogous to 23 by a different methodology,
which will be reported in due course.14
However, in summary, this current route provides a
three-step access to ( )-21–23 for future investigation of
enantioseparation methods.

2094
J. M. Gardiner et al. / Tetrahedron Letters 43 (2002) 2091–2094
reported earlier but with only microanalysis as data:
References
Nassar, R. F.; Issidorides, C. H. J. Org. Chem. 1959, 24,
1832–1833.
7. Piotrowska, H.; Serafin, B.; Urbanski, T. Tetrahedron
1963, 19, 379–383.
8. If the reaction is left for extended times equilibration
occurs to an approximately 1:1 diastereomeric mixture.
9. Kliegel, W.; Preu, L. Chemiker-Zeitung. 1991, 115, 103–
111.
1. Hanessian, S.; Qui, D.; Praphanjan, H.; Reddy, G. V.;
Lou, B. Can. J. Chem. 1996, 74, 1738–1747.
2. (a) Issidoes, C. H.; Galen, R. Org. Synth. 1958, 38,
65–67; (b) Schmidt, M.; Dobner, B.; Nuhn, P. Synlett
2000, 1157–1159.
3. Hall, H. K.; De Blauwe, F.; Yokoyama, Y.; Padias, A.
B. Macromolecules 1980, 13, 252–261.
10. Forbes, D. C.; Ene, D. G.; Doyle, M. P. Synthesis 1998,
879–882.
11. (a) Eliel, E. L. Angew. Chem., Int. Ed. Engl. 1972, 11,
739–750; (b) Eliel, E. L. Acc. Chem. Res. 1970, 3, 1.
12. Juarista, E.; Diaz, F.; Cuellar, G.; Jiminez-Vazquez, H.
A. J. Org. Chem. 1997, 62, 4029–4035.
13. The 2-carbon (n=2) homologation involved (high-yield-
ing) O-allylation, followed by dihydroxylation-periodiate
cleavage, reduction and then silylation.
14. Gardiner, J. M.; Mather, P.; Morjan, R. 2001, unpub-
lished results.
4. These results and all work on pentaerythritol differentia-
tion reported here is taken from: Cooper, M. L. M.Sc.
Thesis UMIST, 1996.
5. (a) 1H NMR shows two identical aryl rings and acetal
protons; 13C shows two types of ring methylenes (axial
or equatorial with respect to the other ring); (b) Jin,
T.-S.; Li, T.-S.; Zhang, Z.-H.; Yuan, Y.-J. Synth. Com-
mun. 1999, 29, 1601–1606.
6. (a) Coxon, A. C.; Stoddart, J. F. J. Chem. Soc., Perkin
Trans. 1 1977, 767–785; (b) Eliel, E. L.; Enanoza, R. M.
J. Am. Chem. Soc. 1972, 94, 8072–8081; (c) Also