Asymmetric double ring-opening of a C2h-symmetric bis-epoxide: Improved enantiomeric excess of the product through enantioselective desymmetrisation and 'proof-reading' steps

Article abstract of DOI:10.1039/b504972e

A new strategy in asymmetric synthesis is described in which the desymmetrisation of a C_2h-symmetric molecule is followed by a subsequent enantioselective 'proof-reading' step. The double asymmetric ring-opening of the bis-epoxide (1R*, 3R*, 5S

Full text of DOI:10.1039/b504972e

View Article Online / Journal Homepage / Table of Contents for this issue
A R T I C L E
Asymmetric double ring-opening of a C2h-symmetric bis-epoxide:
improved enantiomeric excess of the product through enantioselective
desymmetrisation and ‘proof-reading’ steps†
Alan Ironmonger,a,b Peter Stockleyb and Adam Nelson*a,b
a School of Chemistry, University of Leeds, Leeds, UK LS2 9JT
b Astbury Centre for Structural Molecular Biology, University of Leeds, Leeds, UK LS2 9JT
Received 8th April 2005, Accepted 6th May 2005
First published as an Advance Article on the web 23rd May 2005
A new strategy in asymmetric synthesis is described in which the desymmetrisation of a C2h-symmetric molecule is
followed by a subsequent enantioselective ‘proof-reading’ step. The double asymmetric ring-opening of the
bis-epoxide (1R*,3R*,5S*,7S*)-4,8-dioxa-tricyclo[5.1.0.03,5]octane with azidotrimethylsilane, catalysed by a chiral
chromium Salen catalyst, was studied. The reaction involves the initial asymmetric ring-opening of the bis-epoxide to
give the intermediate in moderate enantiomeric excess (ca. 50% ee); the second ring-opening step yields the required
diazido diol, (1S,3S,4S,6S)-4,6-diazidocyclohexane-1,3-diol, in 72% yield and 70% ee. The origin of proof reading
stems from the diversion of the minor enantiomer of the intermediate to a centrosymmetric by-product, a process
which improves the enantiomeric excess of the required product. Using alternative conditions, the reaction was
optimised to yield the required product in >98% ee.
In this generic class of desymmetrisation, careful control of
Introduction
conversion is not necessary for the benefits of the proof-reading
Reactions which involve two asymmetric steps, such as
step to be enjoyed.
sequential1 and parallel2 kinetic resolutions, can offer signif-
icant advantages over their single stepped counterparts. The
enantiomeric excess of the products of many desymmetrisation
Asymmetric double ring-opening of a C2h-symmetric bis-epoxide
reactions may be improved by kinetic resolution, and careful
control of conversion can allow an appropriate balance between
the yield and enantiomeric excess of the desymmetrised product
to be struck.3,4 For example, desymmetrisation of the meso
diacetate 1 using pig liver esterase (PLE) yielded the hydroxy
acetate 2 whose enantiomeric excess could be improved by
selective hydrolysis of its minor enantiomer (Scheme 1).3 Such
an effect may be considered to ‘proof read’ the reaction since the
second enantioselective step serves to improve on the inherent
enantioselectivity of the initial desymmetrisation. Proof-reading
effects in desymmetrisation reactions, especially those which
also raise issues of diastereoselectivity, have been analysed
quantitatively.4
In this paper, we describe the desymmetrisation of a C2h-
symmetric bis-epoxide which is also followed by a subsequent
proof-reading step. However, in marked contrast to the desym-
metrisation of Cs-symmetric molecules such as 1, the second
asymmetric step does not inevitably lead to a meso compound.
Instead, the final product is still chiral and its enantiomeric
excess is influenced by both of the preceding asymmetric steps.
In connection with a research programme directed towards
novel aminoglycoside analogues, we required the diazido diol
5 (Scheme 2). Although a racemic sample had been previously
prepared in high yield,5 we required an asymmetric synthesis
of 5.6 The bis-epoxide 4 was, therefore, treated7 with azi-
dotrimethylsilane and 4 mol% (R,R)-7, and our results are
summarised in Table 1. As has been previously observed,8 the
enantiomeric excess of the product, 5, was optimal at higher
reaction concentrations (compare entries 1 and 2). At lower
concentrations, the required product 5, which had 70% ee, was
accompanied by significant quantities of the centrosymmetric
by-product 6. It is ironic that the centrosymmetric—and hence
achiral9—product 6 was only observed when a chiral catalyst
was used! With sodium azide in buffered water,5 only the racemic
isomer 5, the product of trans-diaxial opening10 of the second
epoxide, had been observed.
We hypothesised that the formation of the centrosymmetric
by-product 6 might have influenced the enantiomeric excess of
the required product 5. The reaction was investigated in more
detail under sub-optimal conditions [initial concentration of 4:
0.5 M in Et2O] in order that the enantiomeric excesses observed
after one and two ring-openings could be easily compared. The
concentrations of the starting material 4, the intermediate and
† Electronic supplementary information (ESI) available: Experimental
procedures and 1H NMR spectra. See http://www.rsc.org/suppdata/
ob/b5/b504972e/
Scheme 1
T h i s j o u r n a l i s
2 3 5 0
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 3 5 0 – 2 3 5 3
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
©

View Article Online
Scheme 2
of the neat unsaturated epoxide 8 with 7.5 mol% (R,R)-7,7
and deprotection, gave the azido alcohol 9 in 57% yield and
80% ee.11 Epoxidation with dimethyldioxirane, generated in situ
from OxoneTM/acetone,12 gave a 50 : 50 diastereomeric mixture
of 10 and 11, which was separated by preparative HPLC;
trimethylsilylation gave the required intermediate 12.
Table 1 Effect of concentration on asymmetric ring opening of bis-
epoxide 4 (see Scheme 2)
4
5
6
Entry Concentration (M; in Et2O) Yielda (%) Eeb (%) Yielda (%)
1
2
3.0
0.5
49c
>98
70
0
70 (72d)
5 (21d)
Fate of the intermediate 12
a Isolated yield. b Determined by chiral analytical HPLC. c In a separate
experiment, the enantiomeric catalyst (S,S)-7 was used; the intermediate
was purified by flash chromatography, and the enantiomeric product ent-
5 was obtained in 68% overall yield, see Supporting Information†. d Yield
in parentheses determined by analysis of the crude product by 500 MHz
1H NMR spectroscopy after 72 h. At this point, starting material had
been completely consumed.
The fate of the major enantiomer of the intermediate was
deduced by treating 12 (80% ee; 0.5 M in Et2O) with azi-
dotrimethylsilane and 4 mol% (R,R)-7 (entry 1, Table 2). The
90 : 10 mixture of 12 and ent-12 yielded a 95 : 5 mixture of 5
and 6 in which 5 had 90% ee, i.e. a 90 : 5 : 5 mixture of 5, 6 and
ent-5. Hence, the second asymmetric step does improve the enan-
tiomeric excess of the intermediate by diverting about half of the
minor enantiomer, ent-12, to the centrosymmetric by-product 6.
In this case, the enantiomeric excess was improved from 80% ee
(in 12) to 90% ee (in 5) by reducing the contamination of the
major enantiomer of 5 with its antipode.
The most obvious strategy for determining the fate of the
minor enantiomer of the intermediate would have been to
subject ent-12 to the same reaction conditions. However, a
more convenient approach, which yielded exactly the same
information, was to subject the same sample of 12 (which had
80% ee) to the reaction catalysed by the enantiomeric catalyst
(S,S)-7 (entry 2, Table 2). In this way, the fate of the minor
enantiomer of the intermediate was deduced unambiguously.
With (S,S)-7, a 90 : 10 mixture of 12 and ent-12 yielded a 60 :
40 mixture of 5 and 6 in which 5 had just 60% ee. In other
words, (R,R)-7 would have converted a 10 : 90 mixture of 12
and ent-12 into a 10 : 40 : 50 mixture of 5, 6 and ent-5. Here,
the same conclusion is reached independently: with (R,R)-7
Fig. 1 Relative concentrations of the starting material 4 (solid triangle),
the intermediate 12 (open triangle), the required product 5 (solid square),
and the centrosymmetric by-product 6 (cross, ×), as a function of time.
the products 5 and 6 were monitored as a function of time by
analysis of aliquots by 500 MHz 1H NMR spectroscopy (Fig. 1).
Asymmetric synthesis of the intermediate 12
Table 2 Fates of enantiomerically enriched samples of the intermedi-
ates 12 and ent-12 in ring-openings catalysed by (R,R)-7
In order to determine the provenance of the centrosymmetric
by-product 6, an enantiomerically enriched sample of the
intermediate 12 was required. Because its concentration was
rather low throughout the course of the asymmetric ring-
opening of 4 (see Fig. 1), an alternative asymmetric synthesis
of 12 was undertaken (Scheme 3). In our hands, treatment
12
5 and 6
Entry Eea(%) 12 : ent-12 5 + ent-5 : 6b,c 5 : 6 : ent-5c 5, eed (%)
1
80
90 : 10
10 : 90
95 : 5
60 : 40
90 : 5 : 5
10 : 40 : 50 60
90
2e
−80d
a Determined by analysis of a sample of 5 by chiral HPLC, prepared
by treatment of 10 with NaN3 in buffered water and deprotection.
b Determined by analysis of the crude reaction mixture by 500 MHz
1H NMR spectroscopy. c The errors in the ratios of products are 5%,
and are rounded accordingly. d Determined by chiral analytical HPLC.
e The fate of the minor enantiomer of 12 was inferred from the reaction
of its enantiomer (which had 80% ee) catalysed by the enantiomeric
catalyst (S,S)-7.
Scheme 3
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 3 5 0 – 2 3 5 3
2 3 5 1

View Article Online
Scheme 4
as the catalyst, about half of ent-12 would be diverted to the
Experimental
centrosymmetric by-product 6.
(1S,3S,4S,6S)-4,6-Diazidocyclohexane-1,3-diol5 5
The proof-reading effect stems from ‘matched’ and ‘mis-
matched’ influences of the substrate and the catalyst (Scheme 4).
After the initial desymmetrisation induced by (R,R)-7, an
enantiomerically enriched intermediate 12 is obtained. For the
major enantiomer of the intermediate, 12, catalyst control-
inducing opening at the R-configured end of the epoxide7—and
substrate control—the preference for trans-diaxial opening10—
are matched, leading to essentially only the major enantiomer of
the product 5. In contrast, substrate and catalyst control are mis-
matched for the minor enantiomer of the intermediate, ent-12,
resulting in some diversion to the alternative centrosymmetric
product, 6. As a result, the enantiomeric excess of the required
product 5 is higher than that of the intermediate 12.
Azidotrimethylsilane (15.5 mL, 117.2 mmol) was added drop-
wise to a stirred solution of the diepoxide13 4 (6.25 g,
55.8 mmol) and (R,R)-N,Nꢀ-bis(3,5-di-tert-butyl-salicylidene)-
1,2-cyclohexane-diaminochromium(III) chloride (4 mol%,
705 mg, 1.12 mmol) in ether (19 mL) and stirred for 96 h.
The reaction mixture was then concentrated under reduced
pressure to give a crude product, which was purified by flash
chromatography eluting with petrol–EtOAc 9 : 1 (+1% Et3N)
to give the diazide as a yellow oil. The TMS-protected diol
was dissolved in 0.05% TFA in MeOH (80 mL) and stirred
for 16 h, evaporated under reduced pressure to give a crude
product, which was purified by flash chromatography eluting
with 8 : 2 petrol–EtOAc to give product which was recrystallised
from CH2Cl2−MeOH as the diol (4.44 g, 49%) as colourless
◦
◦
prisms, mp 98–99 C (from MeOH–CH2Cl2, lit.5 96 C for the
racemate); Rf 0.25 (7 : 3, petrol–EtOAc); [a]2D0 +5.6 (c 1.0 in
CH2Cl2); Found: C, 36.6; H, 5.20; N, 42.3%; C6H10N6O2 requires
C, 36.4; H, 5.10; N, 42.4%); mmax/cm−1 (thin film) 3368, 2923, and
2087; dH (300 MHz; d4-MeOD) 3.71 (2H, q, J 5.8, 1-H and 3-H),
3.45 (2H, q, J 5.8, 4-H and 6-H), 1.80 (2H, t, J 5.8, 2-CH2) and
1.73 (2H, t, J 5.8, 5-CH2); dC (75 MHz; d4-MeOD) 69.8, 63.4,
36.9 and 30.5; m/z (ES−) 197 (100%, M − H).
Effect of proof-reading on the desymmetrisation of the
C2h-symmetric bis-epoxide
The proof-reading effect must also influence the enantiomeric
excess of the product 5 prepared by double asymmetric ring-
opening of the C2h-symmetric epoxide 4. With 0.5 M 4 in ether,
the required product 5 was obtained in 70% ee (see Table 1).
Presumably, the initial desymmetrisation process gave the inter-
mediate 12 in about 50% ee, which was improved by selective
diversion of the minor enantiomer to the centrosymmetric by-
product.
For single step desymmetrisations, relatively large differences
between the activation energies of the competing pathways
(DDG‡) are required to achieve rather modest improvements
in high enantiomeric excesses: for example, an improvement
of about 0.4 kcal mol−1 in DDG‡ is required to increase
the enantiomeric excess of a product from 92% ee to 96%
ee. The proof-reading mechanism described here can yield
similar improvements without requiring an intrinsically more
enantioselective reaction. Although the relative importance of
catalyst and substrate control may be different under the more
concentrated, optimised conditions (3.0 M in ether), it is likely
the observed high enantiomeric excess (>98% ee) of the product
5 derives in part from the proof-reading mechanism available.
The unrecrystallised sample was shown to have >98% ee by
chiral analytical HPLC (Chiracel OD column, 4.6 × 250 mm,
detecting at 225 nm; 95 : 5 hexane–isopropanol; retention times
32 and 35 min).
In a separate experiment, the enantiomeric catalyst (S,S)-7
was used; the intermediate was purified by flash chromatogra-
phy, and the enantiomeric product ent-5 (which had >98% ee)
was obtained in 68% overall yield, see Supporting Information.†
(1R*,2R*,4S*,5S*)-2,5-Diazidocyclohexane-1,4-diol 6 and
NMR reaction monitoring
Azidotrimethylsilane (9.42 mL, 71.4 mmol) was added
dropwise to stirred solution of the diepoxide
4 (2 g,
17.85 mmol) and (R,R)-N,Nꢀ-bis(3,5-di-tert-butyl-salicylidene)-
1,2-cyclohexane-diaminochromium(III) chloride (4 mol%,
451 mg, 0.714 mmol) in ether (36 mL). Samples were taken
at regular intervals and petrol–EtOAc–Et3N (50 : 48 : 2)
was added to these which were concentration under reduced
pressure, filtered through a pipette of silica eluting with petrol–
EtOAc–Et3N (50 : 48 : 2, care was taken as to only remove
the catalyst from the crude reaction mixture), evaporated under
reduced pressure to give a crude product which was examined
Summary
In summary, we have described a novel desymmetrisation strat-
egy which benefits from the cooperation of two enantioselective
steps. The proof-reading enjoyed increased the enantiomeric
excess of the required product 5 considerably. Furthermore,
because the required product was produced in the second of
the enantioselective steps, careful monitoring of the conversion
of the reaction was not necessary.
1
by 500 MHz H NMR spectroscopy. After 96 h, the reaction
mixture was filtered through a short pad of silica as above, and
the crude residue treated with 0.05% TFA in MeOH (50 mL) for
2 3 5 2
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 3 5 0 – 2 3 5 3

View Article Online
30 min. The reaction mixture was concentrated under reduced
pressure to give a crude product, which was recrystallised (twice)
from crude (CH2Cl2–MeOH) and the mother liquor purified by
flash chromatography eluting with 7 : 3 petrol–EtOAc to give
the centrosymmetric diol 6 (183 mg 5%) as colourless plates, mp
158–160 ◦C (from CH2Cl2–MeOH); Rf 0.2 (7 : 3 petrol–EtOAc)
(Found: C, 36.8; H, 5.05; N, 42.3%; C6H10N6O2 requires C, 36.4;
H, 5.10; N, 42.4%); mmax/cm−1 (thin film) 3376, 2939, 2906 and
2105; dH (300 MHz; d4-MeOD) 3.51 (2H, ddd, J 12.5, 9.5 and
4.5, 1-H and 4-H), 3.27 (2H, ddd, J 12.5, 9.5 and 4.5, 2-H and
5-H), 2.03 (2H, app dt, J 12.5 and 4.5, 3-HA and 6-HA) and 1.22
(2H, app q, J 12.5, 3-HB and 6-HB); dC (75 MHz, d4-MeOD)
73.0, 65.3 and 37.7; m/z (ES−) 197 (100%, (M − H); (Found:
[M − H] 197.0789; C6H10N6O2 requires M − H, 197.0792).
Commun., 1990, 567; (d) M. Harding, R. Hodgson and A. Nelson,
J. Chem. Soc., Perkin Trans. 1, 2002, 2403.
2 (a) E. Vedejs and X. Chen, J. Am. Chem. Soc., 1997, 119, 2584; (b) J.
Eames, Angew. Chem., Int. Ed., 2000, 39, 885; (c) E. Vedejs and E.
Rozners, J. Am. Chem. Soc., 2001, 123, 2428.
3 Y.-F. Wang, C.-S. Chen, G. Girdaukas and C. J. Sih, J. Am. Chem.
Soc., 1984, 106, 3695.
4 S. L. Schreiber, T. S. Schreiber and D. B. Smith, J. Am. Chem. Soc.,
1987, 109, 1525.
5 H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem., Int. Ed.,
2001, 40, 2004.
6 The authors point out that, although we have had no problems in
handling azides and solutions of azides, EXTREME CARE should
be taken, especially when heating and handling metal and organic
azides.
7 (a) S. E. Schaus, J. F. Larrow and E. N. Jacobsen, J. Org. Chem.,
1997, 62, 4197; (b) L. E. Martinez, J. L. Leighton, D. H. Carsten and
E. N. Jacobsen, J. Am. Chem. Soc., 1995, 117, 5897.
8 E. N. Jacobsen, Acc. Chem. Res., 2000, 33, 421.
9 M. Anstiss, J. M. Holland, A. Nelson and J. R. Titchmarsh, Synlett,
2003, 1213.
10 J. C. Leffingwell and E. E. Rayals, Tetrahedron Lett., 1965, 3829.
11 The enantiomeric excess of the azido alcohol 9 was determined by
chiral HPLC analysis of its derivative 5. Ring-opening of 10, derived
from 9 (Scheme 3), with NaN3 in buffered water gave the diazido
diol 5.
Acknowledgements
The authors thank the EPSRC Mass Spectrometry Service
Centre, Swansea for mass spectrometry, and James Titchmarsh
for chiral analytical HPLC, and the Wellcome Trust for a prize
studentship.
References
12 (a) R. Curci, M. Fiorentino, L. Troisi, J. O. Edwards and R. H. Pater,
J. Org. Chem., 1980, 45, 4758; (b) P. F. Corey and F. E. Ward, J. Org.
Chem., 1986, 51, 1926.
13 J. Rudolph, K. L. Reddy, J. P. Chiang and K. B. Sharpless, J. Am.
Chem. Soc., 1997, 119, 6189.
1 (a) Z.-W. Guo, S.-H. Wu, C.-S. Chen, G. Girdaukas and C. J. Sih,
J. Am. Chem. Soc., 1990, 112, 4942; (b) W. Kroutil, A. Kleewein
and K. Faber, Tetrahedron: Asymmetry, 1997, 8, 3263; (c) E. L.
Macfarlane, S. M. Roberts and N. Turner, J. Chem. Soc., Chem.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 2 3 5 0 – 2 3 5 3
2 3 5 3