Synthesis of a new and versatile macrocyclic NADH model

Article abstract of DOI:10.1016/S0040-4020(03)00626-4

A new macrocyclic NADH model 1 has been designed and synthesised. The new model consists of the same subunits as previously reported models. However, the present model is designed as such that the pyridine nitrogen of the nicotinamide units are not incorp

Full text of DOI:10.1016/S0040-4020(03)00626-4

Tetrahedron 59 (2003) 4303–4308
Synthesis of a new and versatile macrocyclic NADH model
*
Ulrik Gran
AstraZeneca R&D Molndal, Medicinal Chemistry, S-431 83 Molndal, Sweden
Received 20 January 2003; revised 24 March 2003; accepted 16 April 2003
Abstract—A new macrocyclic NADH model 1 has been designed and synthesised. The new model consists of the same subunits as
previously reported models. However, the present model is designed as such that the pyridine nitrogen of the nicotinamide units are not
incorporated in the macrocyclic framework. Thus, properties such as solubility can easily be varied by alkylation with an appropriate agent.
The macrocyclic framework was prepared in 7 steps. Methylation of the pyridine nitrogens followed by reduction gave the desired model.
This model compound was found to reduce methyl benzoylformate stereoselectively in good yield with 48% enantiomeric excess. q 2003
Elsevier Science Ltd. All rights reserved.
1. Introduction
enantiomeric excess.8 The goal was then to use 2 in a
catalytic cycle by taking advantage of the four nicotinamide
units that are incorporated in the model. By reduction of
only one or two nicotinamide units the model should still be
water soluble and capable of reducing a substrate. The
initial experiments showed that it was not possible to control
the different redox levels of 2, so this route to a water
soluble redox reagent had to be abandoned. Instead,
compound 1 was designed, which has a similar macrocyclic
framework to 2 (Fig. 1). The only difference is that the
dihydropyridine rings are incorporated into the macrocycle
at the 3,5-positions instead of the 1,3-positions. This
modification has several advantages, one being a greater
stability towards basic conditions and nucleophiles. The
bond between the pyridine nitrogen and the methylene
group in 2 is easily broken, thus interrupting the cyclic
structure, and this bond is absent in 1. The modification also
benefits from a greater versatility, since the pyridine
nitrogen is ‘liberated’. Properties such as solubility and
perhaps redox potential can readily be changed by
attaching the appropriate functional groups to the pyridine
nitrogen.
The coenzyme nicotinamide adenine dinucleotide (NADþ/
NADH) is the most common redox reagent in biological
systems, used by hundreds of different enzymes. In the
enzymatic reductions of ketones, one of the two diastereo-
topic hydrogens on the dihydropyridine ring of NADH is
transferred stereoselectively to the substrate.1 The high
performances generally obtained in these reductions have
intrigued organic chemists. Since the initial report by Ohno
and co-workers on asymmetric reduction using an NADH
model,2 there have been many different approaches to
NADH mimicking.3–5 Most of these model reactions have
been studied in acetonitrile. However, hydride transfer is
much faster in more polar solvents such as water.6
Furthermore, very few examples have been concerned
with the catalytic use of NADH models. This is reasonable,
since biomimetic oxidations and reductions usually require
completely different reaction conditions such as solvent. A
water soluble NADH reducing agent could overcome this
problem. Herein is reported the development of a synthetic
route towards water soluble NADH models.
We have previously reported the results from studies of
NADH models of the type represented by 2 (Fig. 1). It was
shown that the oxidised form of 2 was able to form
complexes with derivatives of benzoic acid such as the
dianions of terephthalic and isophthalic acid in water.7 In
these complexes, the guests were intercalated between the
aromatic spacers of 2, showing that the cavity is large
enough to encapsulate a substrate. Compound 2 was also
tested as a reducing agent and was able to reduce methyl
benzoylformate to the corresponding alcohol in 81%
2. Results and discussions
2.1. Synthesis
The synthetic pathway of the previous models was rather
straightforward where nitrogen–carbon bonds were formed
in virtually every step and no modifications prior to
condensation were necessary.8 The synthesis of the new
model compound required a totally different strategy, where
the formation of a carbon–carbon bond resulting in a
methylene group bridging two aromatic rings was the most
challenging task. There are basically two synthetic routes to
Keywords: NADH model; macrocycle; reductions; stereoselectivity.
*
Tel.: þ46-31-7065552; e-mail: ulrik.gran@astrazeneca.com
0040–4020/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4020(03)00626-4

4304
U. Gran / Tetrahedron 59 (2003) 4303–4308
Figure 1. The current NADH model 1 and a previously reported model 2.
macrocycle 1, both using commercially available 5-
bromonicotinic acid as the starting point, as outlined in
Scheme 1.
derivative to an aromatic spacer. The most straightforward
method for the connection is through lithiation followed by
the addition of an electrophile. However, using this method,
an electrophile such as 1,4-bis(halomethyl)benzene has to
be precluded, since it does not mix well with strong bases
such as lithium reagents. Electrophiles containing carbonyl
groups are more suitable in this case, which means that the
methylene groups bridging the pyridine rings and the
aromatic spacer in 4 cannot be obtained in a direct manner.
Furthermore, the metalated analogue of 5-bromonicotinic
acid was a very poor nucleophile. Lithiation of 5-
bromonicotinic acid using 2 equiv. of n-BuLi followed by
addition of an electrophile only gave nicotinic acid after
quenching with water. Quenching the lithiated species using
D2O gave 5-deuterio-nicotinic acid, proving that self-
quenching is not a factor. Converting the acid to the
corresponding pyridyloxazoline and treating it in the same
way did not lead to reaction with an electrophile either. It
seems like the reactivity of lithiated pyridine rings is highly
dependent on its substituents. An electron withdrawing
group on the already electron poor pyridine ring seems to
decrease the reactivity markedly, even with a mildly
Route A was first explored, where the first step was to
condense 2 equiv. of 5-bromonicotinic acid with enantio-
merically pure (1R,2R)-trans-diaminocyclohexane via the
acid chloride prepared in situ, which gave 3 in good yield.
An attempt was then made to use 3, after stannylation, in a
palladium-catalysed Stille coupling. For this purpose, 3 was
converted to the corresponding distannane by lithiation of 3
using n-BuLi followed by addition of tributyltin chloride.
Stille coupling of the distannane with 1,4-bis(bromo-
methyl)benzene would then give the main cyclic structure
of 1. However, the coupling failed and so did attempts with
other benzyl halides. Other strategies to modify 3 in order to
synthesise 1 also failed, either because of low yields or
purification problems. Therefore, route A was abandoned
and route B was pursued instead. The synthesis using route
B is outlined in Scheme 2.
The crucial step was the connection of 2 equiv. of a pyridine
Scheme 1. Two synthetic routes to 1, starting from 5-bromonicotinic acid.

U. Gran / Tetrahedron 59 (2003) 4303–4308
4305
performed at low temperature (2208C) to get the tetraol 7
in good yield. Reoxidation of the terminal alcohols back to
carboxylic acids using KMnO4 as the oxidising agent also
converted the methine alcohols to carbonyl groups to give 8,
which was reduced through Wolff–Kishner reduction to
give the desired diacid 4 in 89% yield from tetraol 7.
The macrocycle 9 could be obtained from 4 and (1R,2R)-
1,2-transdiaminocyclohexane. This could be carried out
either in a ‘one-pot’ procedure using 4 and the diamine in
1:1 stoichiometry, or via compound 10, as outlined in
Scheme 3. The latter procedure gave a higher yield and the
product was also easier to purify. The amide-forming
reactions were initially performed via the acid chloride
prepared in situ, but the use of the coupling reagents TBTU
and HOBt reduced the reactions times considerably.
Alkylation followed by reduction would give the NADH
model 1 (Scheme 4). Since the macrocycle 9 and the by-
products had very low solubility in organic solvents, the
crude product of 9 was alkylated using methyl iodide.
Disappointingly though, the alkylated analogues were not
readily soluble either. However, the iodide counter-ions
could be exchanged for chloride ions using an ion-exchange
resin and the chloride salts were readily soluble in methanol
and water. The crude product was then purified by a similar
method used by Geuder et al. for the purification of
polypyridinium salts.10 Flash chromatography with MeOH/
NH4Cl (aq. 25%) 4:1 as eluent followed by size exclusion
chromatography for the separation of the model compound
from ammonium chloride gave 11. Reduction using basic
sodium dithionite gave the final product 1.
2.2. Reductions
A limited number of reductions were performed using 1 as
the reducing agent. The results are summarised in Table 1
together with results previously obtained for 2.8
The reductions were carried out in methylene chloride using
Mg2þ as a co-catalyst. The results show that NADH model 1
can reduce methyl benzoylformate stereoselectively and in
good yields. The NADH models 1 and 2 seem to be similar
as reducing agents. The same enantiomer (R-mandelate) of
the resulting alcohol is obtained upon reduction and in
similar enantiomeric excess. A more extensive study of the
reduction properties of 2 showed that the selectivity and the
reactivity were dependent on the magnesium ion concen-
tration and the temperature. The enantiomeric excess varied
from 15 to 81% depending on the conditions. These
variations were attributed to a rather flexible structure of
2. Furthermore, the cavity of 2 is large enough to
encapsulate a substrate, which in apolar solvents leads to
poor discrimination between the two faces of the dihydro-
pyridine rings and affects the selectivity negatively. Since
the macrocyclic structure of 1 is similar, it is most likely that
these arguments are valid for 1 as well.
Scheme 2. Reagents and conditions: (a) Et3N, methyl chloroformate,
toluene, rt. (b) LiAlH4, THF, 2788C. (c) Imidazole, TIPS-Cl, DMF, rt. (d)
n-BuLi, terephthaldialdehyde, THF, 21008C. (e) Bu4NF, THF, 2208C. (f)
Na2CO3, KMnO4, H2O, rt. (g) H2NNH2, KOH, ethylene glycol, reflux.
deactivating group such as a carboxylate group. It has
actually been shown that pyridyloxazolines can act as
electrophiles when treated with organolithium or Grignard
reagents.9 Instead, the bromonicotinic acid was reduced and
O-TIPS-protected to give O-TIPS-5-bromopyridine-3-
methanol (5). Lithiation of this species followed by addition
of terephthaldialdehyde worked nicely and gave compound
6 in 65% yield.
3. Conclusions
The diacid can now be obtained from 6 through a series of
transformations. Deprotection using TBAF had to be
A new and versatile NADH model 1 has been synthesised. It
was shown that 1 could function as a stereoselective

4306
U. Gran / Tetrahedron 59 (2003) 4303–4308
Scheme 3. Two pathways for the synthesis of the macrocycle 9.
Table 1. Reductions using 1 and 2
Entry
Substrate
Reducing agent
Ratioa Mg2þ/model
e.e. (%)b
Yield (%)c
1
2
3
4
Methyl benzoylformate
Methyl benzoylformate
Methyl benzoylformate
Methyl benzoylformate
1
2
1
2
1.25
1.25
4
48
54
28
30
100 (2 days)
100 (1 day)
100 (12 h)
100 (4.5 h)
4
The reactions were carried out in methylene chloride at room temperature. Model/substrate 1:1.
a
Magnesium was added in the form of Mg(ClO4)2.
Determined via capillary gas chromatography. The R-isomer was in enantiomeric excess for all entries.
Isolated yields.
b
c
Scheme 4. NADH model 1 is obtained from 9 by alkylation followed by reduction.

U. Gran / Tetrahedron 59 (2003) 4303–4308
4307
reducing agent and the results indicated that it had similar
reduction properties as the previously reported model 2. The
developed synthetic route can be used for preparing NADH
models with different properties such as solubility. Using
the macrocyclic precursor 9 as a starting point, a certain
property can be fine-tuned by alkylation with the appro-
priate agent. An investigation towards a NADH model
which is water soluble both in its reduced and oxidised state
is underway and will be reported in due time.
aqueous layer was extracted with EtOAc (3£50 mL) and
the organic layers were dried (Na2SO4) and filtered. The
solvent was removed under reduced pressure to give crude
5-bromopyridine-3-methanol, which was used in the next
step without purification. To a solution of 5-bromopyridine-
3-methanol (4.15 g, 22.1 mmol) in DMF (20 mL) was
added, under argon, imidazole (3.31 g, 48.6 mmol) followed
by triisopropylsilyl chloride (TIPS-Cl) (5.67 mL,
26.5 mmol). After stirring for 24 h at room temperature
the reaction mixture was evaporated to dryness. Purification
by flash chromatography (silica gel, pentane/EtOAc, 9:1)
1
gave 5 as a colourless oil (6.66 g, 65% overall). H NMR
4. Experimental
4.1. Materials
(CDCl3): d 1.08 (d, 18H, J¼4.3 Hz), 1.19 (m, 3H), 4.84 (s,
2H), 7.86 (s, 1H), 8.50 (s, 1H), 8.57 (s, 1H). 13C NMR: d 11.9,
18.0, 62.2, 136.3,138.8,145.5,149.2,154.0.HRMS:m/zcalcd
for C15H27NSiOBr ([MþH]þ) 344.105, found 344.108.
The reactions were carried out with oven-dried equipment.
Methylene chloride was dried by distillation from calcium
hydride under nitrogen. THF was dried by distillation from
sodium/benzophenone under nitrogen. Dried solvents were
used immediately after distillation. Commercial grade
(peptide-grade) DMF was used. Commercially available
reagents were used without further purification.
4.2.2. Compound 6. To a stirred solution of O-(triisopro-
pylsilyl)-5-bromopyridine-3-methanol (3.52 g, 10.2 mmol)
in anhydrous THF (60 mL) under argon was added n-BuLi
(4.4 mL, 2.45 M in hexane, 10.76 mmol) at 21008C. The
solution was stirred for 15 min and terephthaldialdehyde
(0.76 g, 5.64 mmol) dissolved in anhydrous THF (25 mL)
was added at the same temperature. The mixture was
allowed to slowly warm to room temperature (6 h) before
H2O (60 mL) was added. The aqueous layer was extracted
with EtOAc (3£50 mL) and the combined organic layers
were dried (Na2SO4), filtered and evaporated. Purification
by flash chromatography (silica gel, EtOAc) gave 6 as a
4.2. Methods
Thin-layer chromatography was performed on silica gel
Plate (Merck, silica gel 60 F254). Flash chromatography was
˚
performed using silica gel (Matrex, LC 60 A/35–70 mm).
Size-exclusion chromatography was performed using
Sephadex LH-20 in methanol. 1H (400 MHz)- and 13C
(100.6 MHz) NMR spectra were recorded at 293 K on a
Varian UNITY-400 NMR spectrometer with CDCl3,
DMSO-d6 or methanol-d4 as solvents with tetramethylsilane
as internal standard. IR spectra were recorded on an FT-IR
instrument, Perkin–Elmer 1600. Mass spectra were
recorded on a VG ZabSpec instrument. Positive FAB-MS
(matrix 3-nitrobenzyl alcohol) and positive ESI-MS were
the methods used. Capillary gas chromatography was
performed using a Varian 3300 gas chromatograph
equipped with a silicone fused silica column (column i.d.
0.32 mm, column length 30 m). Optical rotations were
measured on a Perkin–Elmer 241 polarimeter using a 1 dm
cell with a total volume of 1 mL. The enantiomeric excess
was measured using a Hewlett–Packard 5890 gas chro-
matograph equipped with a b-cyclodextrin based column
(Chrompack, CP-Chirasil-dex CB, column i.d. 0.25 mm,
column length 25 m).
1
white solid (2.22 g, 65%). H NMR (CDCl3): d 1.03 (d,
36H, J¼4.3 Hz), 1.13 (m, 6H), 4.79 (s, 4H), 5.77 (s, 2H),
7.27 (s, 4H), 7.67 (s, 2H), 8.16 (s, 2H), 8.39 (s, 2H). 13C
NMR: d 11.9, 18.0, 62.8, 73.5, 127.0, 132.0, 136.7, 139. 1,
143.1, 146.0, 146.2. HRMS: m/z calcd for C38H61N2O4Si2
([MþH]þ) 665.417, found 665.419. IR (KBr): 1107, 1464,
2865, 2942, 3156.
4.2.3. Compound 7. A sample of 6 (2.14 g, 3.22 mmol) was
dissolved in anhydrous THF (35 mL) under argon. Bu4NF
(8.05 mmol, 8.05 mL of a 1.0 M solution in THF) was
added at 2208C and the mixture was stirred at the same
temperature for 9 h. The solvent was evaporated and the
crude product was purified by flash chromatography (silica
gel, EtOAc/methanol, 1:1) to give 7 as a white solid (0.95 g,
84%). 1H NMR (DMSO-d6): d 4.47 (d, 4H, J¼3.5 Hz), 5.30
(tr, 2H, J¼3.5 Hz), 5.74 (s, 2H), 6.02 (s, 2H), 7.34 (s, 4H),
7.66 (s, 2H), 8.34 (s, 2H), 8.45 (s, 2H). 13C NMR: d 60.8,
72.2, 126.3, 132.1, 137.4, 140.6, 144.0, 146.4, 146.8.
HRMS: m/z calcd for C20H21N2O4 ([MþH]þ) 353.150,
found 353.152. IR (KBr): 1027, 1036, 1436, 2858, 3338.
4.2.1. O-(Triisopropylsilyl)-5-bromopyridine-3-metha-
nol (5). To a suspension of 5-bromonicotinic acid (6.0 g,
29.7 mmol) in toluene, Et3N (4.4 mL, 31.5 mmol) was
added at room temperature. After the carboxylic acid was
dissolved, methyl chloroformate (2.42 mL, 31.3 mmol) was
added to the solution and the mixture was stirred for 1.5 h at
room temperature. Precipitated Et3N·HCl was filtered off
and the filtrate was evaporated to dryness to give a mixed
anhydride, which was immediately used in the next step. A
solution of the mixed anhydride in anhydrous THF (50 mL)
was added dropwise to a slurry of LiAlH4 (1.19 g,
31.3 mmol) in anhydrous THF (20 mL) under argon at
2788C and the mixture was stirred for 1.5 h at the same
temperature. It was then poured into 5% aq NaOH solution
(70 mL) and the mixture was passed through celite. The
4.2.4. Compound 8. To a mixture of 7 (0.77 g, 2.21 mmol)
in 10% aq. Na2CO3 solution (11.5 mL) was added KMnO4
(1.66 g, 10.5 mmol) dissolved in water (80 mL). The
mixture was stirred at room temperature for 4 h and was
then passed through celite to eliminate MnO2. The solution
was neutralised with HCl (2 M) to pH 4 and the resulting
precipitate was collected to give 8 as a white solid (0.79 g,
1
95%). H NMR (DMSO-d6): d 7.98 (s, 4H), 8.51 (s, 2H),
9.12 (s, 2H), 9.29 (s, 2H). 13C NMR: d 126.4, 129.2, 129.7,
131.8, 137.8, 139.2, 153.0, 165.3, 193.1. HRMS: m/z calcd
for C20H13N2O6 ([MþH]þ) 377.077, found 377.077. IR
(KBr): 1252, 1661, 1718, 3060.

4308
U. Gran / Tetrahedron 59 (2003) 4303–4308
4.2.5. Compound 4. To a mixture of 8 (1.20 g, 3.19 mmol)
in ethylene glycol (7 mL) was added hydrazine hydrate
(1.55 mL, 31.9 mmol) and KOH (38.3 mmol, 2.15 g). The
solution was heated under reflux for 1.5 h. Volatile material
was distilled off until the internal temperature reached
1808C and the solution was heated again under reflux for
45 min. After cooling water was added and the solution was
neutralised with HCl (2 M) to pH 4. The resulting
precipitate was collected to give 4 as a white solid. The
filtrate was evaporated and a second crop was obtained by
ion-exchange chromatography (Isolute SCX-3) to give
25.8, 32.5, 38.6, 49.3, 55.7, 131.2, 135.6, 137.9, 143.7,
144.2, 144.5, 147.9, 163.3. HRMS: m/z calcd for
C56H64N8O4Cl2 (M2þ) 491.221, found 491.225. IR (KBr):
1560, 1664, 2933, 3386.
4.2.8. Compound 1. Sodium dithionite (182 mg,
1.04 mmol) and sodium carbonate (111 mg, 1.04 mmol)
dissolved in water (2 mL) were added to deaerated solution
of 11 (55 mg, 0.052 mmol) in water (2 mL). The mixture
was stirred for 16 h at room temperature. The solvent was
evaporated and 1 was dissolved in chloroform, filtered to
remove inorganic salts and concentrated to give 1 in 45 mg
1
1.04 g (94%) of 4 as a white solid. H NMR (DMSO-d6):
1
d 4.01 (s, 4H), 7.22 (s, 4H), 8.05 (s, 2H), 8.68 (s, 2H), 8.88
(s, 2H). 13C NMR: d 37.4, 127.3, 129.3, 136.8, 137.1, 138.4,
148.3, 153.2, 166.6. HRMS: m/z calcd for C20H17N2O4
([MþH]þ) 349.119, found 349.119. IR (KBr): 1208, 1716,
3056.
(95%) as a red solid. H NMR (CDCl3): d 1.3 (m, broad,
4H), 1.73 (m, broad, 4H), 1.81 (m, broad, 4H), 2.11 (m,
broad, 4H), 2.88 (s, 12H), 2.90 (m(overlapping), 8H), 3.06
(d, 4H, J¼9.5 Hz), 3.16 (d, 4H, J¼9.5 Hz), 3.65 (m, 4H),
5.48 (s, 4H), 5.87 (d, broad, 4H), 6.75 (s, 4H), 7.13 (s, 8H).
4.2.6. Compound 10. To a stirred mixture of 4 (85 mg,
0.24 mmol) in DMF (20 mL) under argon was added TBTU
(172 mg, 0.54 mmol) and HOBt (66 mg, 0.48 mmol). After
the dicarboxylic acid was dissolved, (1R,2R)-(2)-1,2-
transdiaminocyclohexane (167 mg, 1.47 mmol) was added
and the solution was stirred over night at room temperature.
The solvent was evaporated and the crude product was
purified by flash chromatography (silica gel, gradient from
EtOAc/methanol 1:1 to methanol/Et3N, 99:1), which gave
10 in 84 mg (64%) as a white solid. 1H NMR (methanol-d4):
d 1.33 (m, broad, 8H), 1.75 (m, broad, 4H), 1.97 (m, broad,
4H), 2.64 (m, 2H), 3.71 (m, 2H), 4.04 (s, 4H), 7.20 (s, 4H),
8.10 (s, 2H), 8.54 (s, 2H), 8.83 (s, 2H). 13C NMR: d 26.0,
26.2, 33.0, 35.3, 39.1, 55.1, 57.4, 130.4, 132.1, 137.3, 138.9,
139.4, 147.0, 152.7, 168.0. HRMS: m/z calcd for
C32H41N6O2 ([MþH]þ) 541.329, found 541.327. IR
(KBr): 1326, 1543, 1642, 2931, 3278.
4.3. General procedure for reductions
A 10 mL round-bottomed flask was charged with mag-
nesium perchlorate and the model compound (80–
100 mmol), sealed with a rubber septum, and flushed with
argon. Freshly distilled methylene chloride (2.5–3 mL) was
added via syringe and the mixture was stirred for 1–2 h,
whereupon the substrate was added via syringe. The flask
was kept in the dark and the reaction mixture was stirred
under argon. The reaction was quenched by adding 7–8 mL
of water. The aqueous phase was extracted with methylene
chloride (3£10 mL) and the combined organic phases were
dried over Na2SO4, filtered and concentrated. The optical
rotation was measured to determine the enantiomer in
excess and the enantiomeric excess was measured using
capillary GC.
4.2.7. Compound 11. To a stirred mixture of 4 (122 mg,
0.35 mmol) in DMF (100 mL) under argon was added
TBTU (247 mg, 0.77 mmol) and HOBt (107 mg,
0.70 mmol). After the dicarboxylic acid was dissolved, a
sample of 10 (189 mg, 0.35 mmol) dissolved in DMF
(15 mL) was added and the solution was stirred over night at
room temperature. The solvent was evaporated and
acetonitrile (40 mL) was added. The resulting precipitate
was collected to give crude 9, which was used in the next
step without purification. To a stirred mixture of 9 in DMF
(5 mL) was methyl iodide (3 mL) and the mixture was
stirred under argon for seven days at room temperature. The
solvent was evaporated and the resulting iodide salt was
converted to the chloride salt using ion-exchange chroma-
tography (Amberlite IRA-400 (Cl)) with methanol as eluent.
The solvent was evaporated and the crude product was
purified by flash chromatography (silica gel, methanol/
NH4Cl (25% aq.), 4:1) followed by size exclusion
chromatography with methanol as eluent to give 11 in
References
1. Dugas, H. Bioorganic Chemistry. A Chemical Approach to
Enzyme Action; 2nded. Springer: New York, 1989; pp 187–189.
2. Ohnishi, Y.; Kagami, M.; Ohno, A. J. Am. Chem. Soc. 1975,
97, 4766–4768.
3. Burgess, V. A.; Davies, S. G.; Skerlj, R. T. Tetrahedron:
Asymmetry 1991, 2, 299–328.
´
4. Dupas, G.; Levacher, V.; Bourgignon, J.; Queguiner, G.
Heterocycles 1994, 39, 405–429.
5. Murakami, Y.; Kikuchi, J.; Hisaeda, Y.; Hayashida, O. Chem.
Rev. 1996, 96, 721–758.
6. von Eikeren, P.; Grier, D. L. J. Am. Chem. Soc. 1976, 98,
4655–4657.
¨
7. Gran, U.; Wennerstrom, O.; Westman, G. Tetrahedron 2001,
57, 8897–8902.
8. Gran, U.; Wennerstrom, O.; Westman, G. Tetrahedron:
¨
Asymmetry 2000, 11, 3027–3040.
9. Meyers, A. I.; Gabel, R. A. J. Org. Chem. 1982, 47,
2633–2637.
1
92 mg (25%). H NMR (methanol-d4): d 1.43 (m, broad,
4H), 1.66 (m, broad, 4H), 1.87 (m, broad, 4H), 2.05 (m,
broad, 4H), 4.02 (m, 4H) 4.18 (s, 8H), 4.41 (s, 12H), 7.30 (s,
8H), 8.52 (s, 4H), 9.00 (s, 4H), 9.03 (s, 4H). 13C NMR: d
10. Geuder, W.; Hu¨nig, S.; Suchy, A. Tetrahedron 1986, 42,
1665–1677.