First enantioselective non-biological synthesis of asymmetrised tris(hydroxymethyl)methane (THYM*) and bis(hydroxymethyl)acetaldehyde (BHYMA*)

Article abstract of DOI:10.1016/S0040-4039(01)01048-6

An asymmetric synthesis of a chiral non-racemic (O-benzyl, O′-silyl) derivative of the latent C_3v-symmetric tris(hydroxymethyl)methane (THYM*) and of the bis(hydroxymethyl)acetaldehyde (BHYMA*) in 6 steps, 38% overall yield and 7 steps, 36% ove

Full text of DOI:10.1016/S0040-4039(01)01048-6

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 5421–5424
First enantioselective non-biological synthesis of asymmetrised
tris(hydroxymethyl)methane (THYM*) and
bis(hydroxymethyl)acetaldehyde (BHYMA*)†
Irene Izzo, Matteo Scioscia, Pasquale Del Gaudio and Francesco De Riccardis*
Dipartimento di Chimica, Universita` di Salerno, Via S. Allende, Baronissi-I-84081, Salerno, Italy
Received 28 May 2001; accepted 14 June 2001
Abstract—An asymmetric synthesis of a chiral non-racemic (O-benzyl, O%-silyl) derivative of the latent C3v-symmetric tris(hydroxy-
methyl)methane (THYM*) and of the bis(hydroxymethyl)acetaldehyde (BHYMA*) in 6 steps, 38% overall yield and 7 steps, 36%
overall yield, respectively, is described starting from the commercially available 4-nitrobenzoate derivative of 17. The method
involves the Sharpless asymmetric epoxidation and a regioselective copper-mediated oxirane ring opening, as key steps. © 2001
Elsevier Science Ltd. All rights reserved.
In the context of the synthesis of new chiral non-
racemic building blocks, the asymmetrised tris(hydroxy-
methyl)methane (THYM*, 1) and bis(hydroxy-
methyl)acetaldehyde (BHYMA*, 2) appear to be the
ideal precursors to most tertiary C-branched stereo-
genic centres because of the versatility of their oxy-
genated functionalities.1
In this paper we wish to report the first, facile, high
yielding, non-enzymatic enantioselective methodology
for the synthesis of both enantiomers of 2-[(benzyl-
oxy)methyl]-3-(tert-butyldiphenylsilyloxy)-1-propanol
(3)4 and 2-[(benzyloxy)methyl]-3-(tert-butyldiphenylsil-
yloxy)-propanal (4), using the Sharpless asymmetric
epoxidation as the key step. The benzyl and tert-
butyldiphenylsilyl protective groups were chosen for
their ideal orthogonal deprotective modes.5
OH
O
R2-O
2, R1 R2
1, R1 R2
R2-O
H
H
=
=
In a first attempt, we reasoned that the simplest
approach to
a (O-benzyl, O%-silyl) derivative of
O-R1
O-R1
BHYMA* 2 and, consequently, to the related THYM*
1, would have been the stereospecific methylaluminum
(S)-3, R1 = Bn, R2 = TPS;
(R)-3, R1 = TPS, R2 = Bn
(S)-4, R1 = Bn, R2 = TPS;
(R)-4, R1 = TPS, R2 = Bn
bis(4-bromo-2,6-di-tert-butylphenoxide)
(MABR)-
catalysed rearrangement6 of the readily available opti-
cally active epoxysilyl ether 5. This Lewis acid-catalysed
rearrangement of epoxides can give, in principle, the
The preparation of asymmetric derivatives of 1 and 2,
mainly through enzymatic desymmetrisation of suitable
diols and diacetates,2 followed by protective and func-
tional groups manipulation, and their use in organic
synthesis has been recently summarised in a compre-
hensive review by Banfi and Guanti.3 Reported yields
and enantioselectivities are slightly beyond 30 and 98%,
respectively.
optically
active
(S)-2-[(benzyloxy)methyl]-3-(tert-
butyldimethylsilyloxy)-propanal (6) in a single step.
O
O
Bn-O
MABR
H
H
TBS-O
CH2Cl2
O-TBS
H
O-Bn
6
5
Consequently, according to Scheme 1, (2S,3S)-4-(ben-
zyloxy)-2,3-epoxybutan-1-ol (9)7 was synthesised from
commercially available (Z)-4-(benzyloxy)-2-buten-1-ol
(7) and almost quantitatively silylated to give 5 in 73%
overall yield (e.e. 98%).8
Keywords: asymmetric synthesis; epoxidation; chiral building blocks.
* Corresponding author.
† Dedicated to the memory of Professor Guido Sodano, deceased
June 7th, 2001.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01048-6

5422
I. Izzo et al. / Tetrahedron Letters 42 (2001) 5421–5424
O
a
c
b
Bn-O
H
BnO
5
BnO
OH
OH
OH
H
7
8
9
Scheme 1. (a) (i) 1.2 equiv. of pyridinium chlorochromate, Celite® (0.3 g/mmol), CH2Cl2, rt, 18 h; (ii) 1.4 equiv. of NaBH4, MeOH
−40°C, 2 h, 89%; (b) 1.0 equiv. of Ti(Oi-Pr)4, 1.1 equiv. of (+)-DET, 2.0 equiv. of t-BuOOH, CH2Cl2, −18°C, 18 h, 84%, (e.e.
98%); (c) 1.5 equiv. of TBDMS-Cl, 2 equiv. of imidazole, DMF, 0°C, 2 h, 98%.
Unfortunately, the organoaluminum-promoted rear-
rangement of 5 failed to give the expected b-branched
aldehyde 6. Even forcing the reaction conditions with
the use of higher reaction temperatures (though still at
or below 0°C) and increasing the equivalents of MABR
(from 0.2 to 1.0 or 2.0 equiv.) we recovered the unal-
tered starting material.9
temperature the reaction proceeded very slowly while
higher concentration resulted in lower regioselectivity.
The minor, undesired, 1,2-diol was oxidatively cleaved
with sodium periodate14 to facilitate the purification of
the desired 1,3-diol 13.
Regioselective silylation with tert-butyl-diphenylsilyl
chloride and a nonreductive debenzylation with the
mild Lewis acid BCl3·SMe2 complex,15 furnished the
(2R,3R)-4-(tert-butyldiphenylsilyloxy)-3-ethenyl-butan-
1,2-diol (15).
In consideration of the stability of 5 to MABR, we
turned our attention to the supposed more reactive
substrate 11. The parent epoxide 10 was enantiosec-
tively prepared in 95% e.e.8 following a known proce-
dure.10 Standard silylation and stoichiometric11
Yamamoto’s rearrangement6 (Scheme 2) on silyl ether
11 induced an unexpected extended rearrangement
affording an unstable adduct which was tentatively
assigned as 12.12
C-1 sacrifical oxidative cleavage, with NaIO4 in H2O/
MeOH, followed by in situ reduction of the aldehyde
16
with NaBH4 afforded the key intermediate (S)-2-
[(tert-butyldiphenylsilyloxy)methyl]-but-3-en-1-ol (16)
in good overall yield.
In view of these discouraging results we turned our
attention back to the trans-epoxide 9. This time our
plan relied on a regioselective copper-mediated oxirane
ring opening,13 using a vinyl moiety as a C1-masked
unit. Thus, stereospecific opening of the oxirane 9 with
vinylmagnesium bromide in the presence of cuprous
iodide led to the desired monoprotected (2R,3R)-4-
(benzyloxy)-2-ethenyl-butan-1,3-diol (13) and its
regioisomer in a 83:17 ratio (Scheme 3). Control of the
reaction temperature, choice of the optimal reagents’
concentration (0.05 M, for 9) and solvents ratio were
crucial to the success of the ring opening, since at low
Similar results were obtained starting directly from the
cis-epoxide 1717 (Scheme 4).
The two reaction sequences reported in Schemes 3 and
4 show that yields and enantiopurities were higher
starting from the cis-epoxide 17 (52% overall yield,
e.e.>98%8), compared with those starting from the
trans-epoxide 9 (41% overall yield, e.e.=98%8). More-
over the preparation of the trans-epoxide 9 requires the
preliminary isomerisation of the (Z)-(benzyloxy)-2-
buten-1-ol (7, see Scheme 1).
3
1
O
t-Bu
Br
O
2
H
O-TBS
AlL2
O
b
t-Bu
O-R
H
OH
Br
H
t-Bu
O
10, R = H
11, R = TBS
t-Bu
O-TBS
a
H
12
Scheme 2. (a) 1.5 equiv. of TBDMS-Cl, 2 equiv. of imidazole, DMF, 0°C, 2 h, 67%; (b) 0.2 equiv. of MABR, CH2Cl2, −78°C,
0.6 h.
OH
OH
a
b
d
3
R-O
HO
OTBDPS
1
2
Bn-O
9
2
3
1
O-TBDPS
OH
16
R = Bn, 14
R = H, 15
13
c
Scheme 3. (a) (i) 3 equiv. of CH2ꢀCH-MgBr, 0.4 equiv. of CuI, Et2O/THF (5:1), −10°C, 18 h; (ii) 1.0 equiv. of NaIO4, THF/H2O
(1:1) 0°C, 4 h, 57%; (b) 1.5 equiv. of TBDPS-Cl, 1.0 equiv. of DMAP, CH2Cl2/pyridine (20:1), 0°C, 18 h, 96%; (c) 7 equiv. of
BCl3·SMe2, CH2Cl2, 0°C, 2.5 h, 82%; (d) (i) 3.0 equiv. of NaIO4, MeOH/H2O (1:1), 0°C, 3 h; (ii) 3.0 equiv. of NaBH4, MeOH,
0°C, 3 h, 92%.

I. Izzo et al. / Tetrahedron Letters 42 (2001) 5421–5424
5423
OH
O
R2-O
H
a
H
d
16
O-R1
Bn-O
OH
17
R1 = Bn, R2 = H; 18
R1 = Bn, R2 = TBDPS; 19
R1 = H, R2 = TBDPS; 20
b
c
Scheme 4. (a) (i) 3 equiv. of CH2ꢀCH-MgBr, 0.4 equiv. of CuI, Et2O/THF (5:1), −10°C, 18 h; (ii) 1.0 equiv. of NaIO4, THF/H2O
(1:1) 0°C, 4 h, 63%; (b) 1.5 equiv. of TBDPS-Cl, 1.0 equiv. of DMAP, CH2Cl2/pyridine (20:1), 0°C, 18 h, 96%; (c) 7 equiv. of
BCl3·SMe2, CH2Cl2, 0°C, 2.5 h, 97%; (d) (i) 3.0 equiv. of NaIO4, MeOH/H2O (1:1), 0°C, 3 h; (ii) 3.0 equiv. of NaBH4, MeOH,
0°C, 3 h, 88%.
b
a
c
Bn-O
O-TBDPS
(S)-3
(S)-4
16
21
Scheme 5. (a) 1.3 equiv. of BnBr, 1.1 equiv. of Ag2O, CH2Cl2, rt, 48 h, 80%; (b) (i) O3, MeOH/CH2Cl2 (1.7:1), −78°C, 0.4 h; (ii)
,
Me2S, 5 min; (iii) 8 equiv. of NaBH4, −78°C“rt, 18 h, 93%; (c) 1.8 equiv. PDC, 4 A ms, CH2Cl2, 95%.
Finally, protection of the free hydroxy group with
benzyl bromide in presence of Ag2O18 and reductive
ozonolysis with Me2S/NaBH419 furnished (S)-3-(benzyl-
oxy)-2-[(tert-butyldiphenylsilyloxy)methyl]-propan-1-ol
(3) in 39% overall yield from 17 and e.e.>98%8 (Scheme
5).20 PDC oxidation21 on (S)-3 synthesised from the
cis-epoxide 17 gave the expected (2S)-3-(benzyloxy)-2-
6. Maruoka, K.; Ooi, T.; Nagahara, S.; Yamamoto, H.
Tetrahedron 1991, 47, 6983–6998.
7. (a) Katsuki, T.; Lee, A. W. M.; Ma, P.; Martin, V. S.;
Masamune, S.; Sharpless, K. B.; Tuddenham, D.;
Walker, F. J. J. Org. Chem. 1982, 47, 1378–1380; (b)
Lo¨fstedt, J.; Pettersson-Fasth, H.; Ba¨ckvall, J.-E. Tetra-
hedron 2000, 56, 2225–2230; (c) For the synthesis of 8 it
was followed the procedure in: Ko, S. Y.; Lee, A. W. M.;
Masamune, S.; Reed, L. A.; Sharpless, K. B.; Walker, F.
J. Tetrahedron 1990, 46, 245–264.
[(tert-butyldiphenylsilyloxy)methyl]-propanal
(S)-4
without any concurrent racemisation.22
In summary, we have reported the first enantioselective
non-biological syntheses of stereodefined asymmetrised
tris(hydroxymethyl)methane (THYM*) and bis(hy-
droxymethyl)acetaldehyde (BHYMA*). These methods
provide a new straightforward access to their (O-ben-
zyl, O%-silyl) derivatives with good chemical yields and
excellent enantioselectivities.
8. The enantiomeric purity was checked through 400 MHz
1H NMR Mosher’s esters analyses (Dale, J. A.; Mosher,
H. S. J. Am. Chem. Soc. 1973, 95, 512–519) using C6D6
as solvent. These were carried out synthesising the two
diastereomeric derivatives using (R)-(−)- and (S)-(+)-a-
methoxy-a-[(trifluoromethyl)phenylacetyl]chlorides ((R)-
(−)- and (S)-(+)-MTPA-Cl) and integrating the relevant
1
peaks in the H NMR spectrum.
9. The organoaluminum-promoted rearrangement proceeds
through an anti migration of the C-1 silyloxymethyl
group towards the C-3 of the epoxide. Our result seems
to confirm that silylated 2,3-epoxy alcohols not stabilising
a transient positive charge on C-3 do not undergo the
Yamamoto’s rearrangement (see Ref. 6).
10. Wershofen, S.; Sharf, H.-D. Synthesis 1988, 854–858.
11. No reaction was observed when the Yamamoto’s rear-
rangement was performed with catalytic amounts of
MABR.
Acknowledgements
This work has been supported by the MURST (PRIN
‘Chimica dei Composti Organici di Interesse
Biologico’).
References
12. Instability of 12 prevented full spectroscopic characterisa-
1. Larock, R. Comprehensive Organic Transformation;
Wiley-VCH: New York, 1999.
1
tion. The structure is consistent with the H NMR spec-
tral data and homonuclear decoupling experiments. 12:
1H NMR (400 MHz, CDCl3): l 0.07 (6 H, s,-Si(CH3)2),
0.90 (9 H, s, -SiC(CH3)3), 1.24 (18 H, s, -C(CH3)3), 2.86
(2 H, m, H-5 and H%-5), 3.34 (1 H, dd, J=9.8, 8.8 Hz,
H-1), 3.55 (1 H, dd, J=9.8, 3.5 Hz, H%-1), 4.11 (1 H, m,
H-2), 5.52 (1 H, dd, J=15.6, 5.8 Hz, H-3), 5.60 (1 H, m,
H-4), 6.72 (2 H, s, -C=CH-).
2. The only exception is the non-biological diastereoselective
preparation reported by Harada et al.: Harada, T.;
Hayashiya, T.; Wade, I.; Iwa-ake, N.; Oku, A. J. Am.
Chem. Soc. 1987, 109, 527–532; Harada, T.; Oku, A.
Synlett 1994, 95–104.
3. Banfi, L.; Guanti, G. Eur. J. Org. Chem. 1998, 745–757.
4. Satisfactory analytical data (1H and 13C NMR, EI-MS
spectra and elemental analyses) were obtained for all the
new compounds.
13. Tius, M.; Fauq, A. H. J. Org. Chem. 1983, 48, 4131–
4232.
14. Richardson, T. I.; Rychnovsky, S. D. J. Org. Chem. 1996,
62, 4219–4231.
5. Kocienski, P. J. Protecting Groups; Georg Thieme: Stutt-
gart, 1994.

5424
I. Izzo et al. / Tetrahedron Letters 42 (2001) 5421–5424
15. Congreve, M. S.; Davison, E. C.; Furhy, M. A. M.;
Holmes, A. B.; Payne, A. N.; Robinson, R. A.; Ward, S.
E. Synlett 1993, 663–664.
dd, J=9.0, 6.3 Hz, -CHHOBn), 3.72 (1H, dd, J=9.7, 6.1
Hz, -CHHOTBDPS), 3.75 (1H, dd, J=9.7, 4.4 Hz,
-CHHOTBDPS), 4.50 (2H, s, -CH2Ph), 5.11 (1H, dd,
J=10.8, 1.0 Hz, CHHꢀCH-), 5.12 (1H, dd, J=17.7, 1.0
16. Lee, M. G.; Du, F. J.; Chun, M. W.; Chu, C. K. J. Org.
Chem. 1997, 62, 1991–1995.
Hz, CHHꢀCH-), 5.84 (1 H, m, CH2ꢀCH-), 7.30–7.45
13
17. (2S,3R)-(+)-3-Benzyloxy-2,3-epoxy-1-butyl
4-nitroben-
(11H, m, C6H5-), 7.64 (4H, m, C6H5-);
C NMR (100
zoate and its enantiomer are available from Fluka
Chemie AG, Switzerland (e.e.>98%). Hydrolysis of the
p-nitrobenzoate ester with 1% NaOH in MeOH gave the
desired epoxy alcohol 17 in quantitative yield. See: Pet-
tersson-Fasth, H.; Ba¨ckvall, J.-E. J. Org. Chem. 1995, 60,
6091–6096. Epoxide 17 can also be synthesised through
Sharpless asymmetric epoxidation7a in 84% yield and 90%
MHz, CDCl3): l 19.3, 26.8 (×3), 46.2, 64.1, 70.3, 73.1,
116.6, 127.4 (×7), 128.4 (×2), 129.5 (×2), 133.7 (×2), 135.6
(×4), 137.3, 138.5; HR EIMS m/z 430.2317 (calcd
430.2328 for C28H34O2Si).
(S)-3: [h]D −3.2 (c=1.0, CHCl3); 1H NMR (400 MHz,
CDCl3): l 1.08 (9H, s, (CH3)3C-Si), 2.14 (1H, m, H-2),
2.68 (1H, bs, -OH), 3.62 (1H, dd, J=9.3, 6.7 Hz,
-CHHOBn), 3.66 (1H, dd, J=9.3, 5.9 Hz, -CHHOBn),
3.80–3.85 (4H, m, -CH2OTBDPS and -CH2OH), 4.51
(2H, s, -CH2Ph), 7.30–7.45 (11H, m, C6H5-), 7.68 (4H, m,
C6H5-). 13C NMR (100 MHz, CDCl3): l 19.1, 26.8 (×3),
43.2, 63.4, 64.0, 70.0, 73.3, 127.6 (×7), 128.3 (×2), 129.7
(×2), 133.2 (×2), 135.5 (×4), 138.0; HR EIMS m/z
434.2265 (calcd 434.2277 for C27H34O3Si).
1
e.e. (Mosher’s esters H NMR analyses).8 It is possible to
improve its enantiomeric purity (e.e.>98%) through
recrystallisation of its 4-nitrobenzoate and subsequent
hydrolysis.
18. Van Hijfte, L.; Little, R. D. J. Org. Chem. 1985, 50,
3942–3943. Compound 21 is often contaminated by a
30–40% of the bis-benzyl ether.
19. Ozonolysis performed on batches containing less than
40–50 mg of 21 resulted in lower yields. When small
amounts of 4 are requested, it is more convenient to
obtain it by oxidation of the parent alcohol. Banfi, L.;
Guanti, G.; Narisano, E. Tetrahedron 1993, 49, 7385–
7392.
(S)-4: [h]D 8.3 (c=0.8, CHCl3); 1H NMR (400 MHz,
CDCl3): l 1.05 (9H, s, (CH3)3C-Si), 2.78 (1H, m, H-2),
3.82 (1H, dd, J=9.4, 6.2 Hz, -CHHOBn), 3.90 (1H, dd,
J=9.4, 5.9 Hz, -CHHOBn), 4.04 (1H, dd, J=10.4, 2.6
Hz, -CHHOTBDPS), 4.07 (1H, dd, J=10.4, 5.4 Hz,
-CHHOTBDPS), 4.47 (2H, s, -CH2Ph), 7.30–7.45 (11H,
m, C6H5-), 7.66 (4H, m, C6H5-); 13C NMR (100 MHz,
CDCl3): l 19.1, 26.8 (×3), 54.5, 60.1, 65.8, 73.3, 127.5
(×3), 127.7 (×4), 128.3 (×2), 129.7 (×2), 133.0 (×2), 135.5
(×4), 138.0, 202.6; HR EIMS m/z 432.2117 (calcd
432.2121 for C27H32O3Si).
20. Selected data for compounds are as follows: 16: [h]D
−17.5 (c=1.0, CHCl3); 1H NMR (400 MHz, CDCl3): l
1.07 (9H, s, (CH3)3C-Si), 2.27 (1H,bs, -OH), 2.55 (1H, m,
H-2), 3.70–3.83 (4 H, m, -CH2OH and -CH2OTBDPS),
5.11 (1 H, dd, J=17.7, 1.0 Hz, CHHꢀCH-), 5.13 (1H, dd,
J=10.3, 1.0 Hz, CHHꢀCH-), 5.69 (1H, m, CH2ꢀCH-),
13
21. Herscovici, J.; Antonakis, K. J. Chem. Soc. Chem.
Comm. 1980, 561–562.
7.43 (6H, m, C6H5-), 7.68 (4H, m, C6H5-);
C NMR
(100 MHz, CDCl3): l 19.1, 26.8 (×3), 47.6, 64.7, 66.0,
117.6, 127.7 (×4), 129.8 (×2), 133.1 (×2), 135.5 (×4),
135.8; HR EIMS m/z 340.1859 (calcd 340.1837 for
C21H28O2Si).
1
22. 400 MHz H NMR Mosher’s esters analysis, using C6D6
as solvent, on the alcohol derived from the NaBH4
reduction of (S)-4, showed that no racemisation occurred
during the oxidation and/or the work-up processes. The
only other method known that avoids racemisation is a
modification of the Swern oxidation.19
1
Compound 21: [h]D −5.1 (c=1.0, CHCl3); H NMR (400
MHz, CDCl3): l 1.03 (9H, s, (CH3)3C-Si), 2.58 (1H, m,
H-2), 3.55 (1H, dd, J=9.0, 6.1 Hz, -CHHOBn), 3.65 (1H,
.