The elusive 8-fluoroadenosine: a simple non-enzymatic synthesis and characterization

Article abstract of DOI:10.1016/j.tet.2007.02.053

The only successful synthesis of 8-fluoroadenosine reported until now relied on an enzymatic removal of the acetate protecting groups using thermally resistant hydrolases. In the present communication we describe the first non-enzymatic synthesis of 8-fluoroadenosine. According to this, the C₈-fluorine atom was introduced in a halogen-exchange process performed at elevated temperature. The chief obstacle in the synthesis of 8-fluoroadenosine, the removal of the protecting groups in the presence of the labile C₈-F bond, was addressed by judicious choice of acid-labile protecting groups. Their deprotection in the presence of C₈-F is described. Using this newly developed procedure, significant quantities of 8-fluoroadenosine were synthesized and, for the first time, its physicochemical properties including pH-dependent stability, examined in detail. The intermediate generation of 8-fluoroadenosines as a tool to increase the reaction rates of nucleophilic substitutions was briefly examined and successfully demonstrated with the example of 8-cyanoadenosine. The presented procedure is applicable to the synthesis of various adenosine analogs with potential pharmacological significance.

Full text of DOI:10.1016/j.tet.2007.02.053

Tetrahedron 63 (2007) 3782–3789
The elusive 8-fluoroadenosine: a simple non-enzymatic
synthesis and characterization
*
Gabor Butora, Christoph Schmitt, Dorothy A. Levorse, Eric Streckfuss,
George A. Doss and Malcolm MacCoss
Department of Medicinal Chemistry, Merck Research Laboratories, Rahway, NJ 07065, USA
Received 15 December 2006; revised 12 February 2007; accepted 14 February 2007
Available online 20 February 2007
Abstract—The only successful synthesis of 8-fluoroadenosine reported until now relied on an enzymatic removal of the acetate protecting
groups using thermally resistant hydrolases. In the present communication we describe the first non-enzymatic synthesis of 8-fluoroadenosine.
According to this, the C8-fluorine atom was introduced in a halogen-exchange process performed at elevated temperature. The chief obstacle
in the synthesis of 8-fluoroadenosine, the removal of the protecting groups in the presence of the labile C8–F bond, was addressed by judicious
choice of acid-labile protecting groups. Their deprotection in the presence of C8–F is described. Using this newly developed procedure,
significant quantities of 8-fluoroadenosine were synthesized and, for the first time, its physicochemical properties including pH-dependent
stability, examined in detail. The intermediate generation of 8-fluoroadenosines as a tool to increase the reaction rates of nucleophilic
substitutions was briefly examined and successfully demonstrated with the example of 8-cyanoadenosine. The presented procedure is
applicable to the synthesis of various adenosine analogs with potential pharmacological significance.
Ó 2007 Elsevier Ltd. All rights reserved.
1. Introduction
8-Bromoadenosine (1d) can be easily synthesized on a large
scale by direct bromination of adenosine (1a) in a buffered
medium.15,16 Synthesis of 8-iodoadenosine (1e) is more
problematic and requires a lithiation prior to quenching
with molecular iodine.17 8-Chloroadenosine (1c) cannot be
obtained by direct chlorination at all, and alternative
methods had to be developed.18,19
8-Halogen substituted adenosines can serve as a convenient
starting point in the syntheses of unnatural purine nucleo-
sides.1,2 The most versatile of them, 8-bromoadenosine3
(1d, Fig. 1), first described in 1964, is often used in reactions
involving direct displacements with nucleophiles4,5 as well
as transition metal catalyzed cross-couplings6,7. The syn-
thetic utility of 8-chloroadenosine8 (1c) and 8-iodoadeno-
sine9 (1e), first synthesized in 1972 and 1971, respectively,
is much more limited.10–12 In spite of numerous attempts,13
the first successful synthesis of 8-fluoroadenosine14 (1b) was
not accomplished until 1998.
A purported synthesis9,20 of 8-fluoroadenosine (1b) involved
a Balz–Schiemann reaction of 20,30,50-tri-O-acetyl-8-amino-
adenosine (1f) followed by deprotection with methanolic
ammonia that appeared to be, in retrospect, too harsh to pro-
duce the desired 8-fluoroadenosine (1b). In fact numerous
attempts at synthesis of acetylated 8-fluoroadenosines in
our laboratory using various modifications of the Balz–
Schiemann reaction failed to afford detectable (LCMS)
amounts of this product. Similar conclusions were drawn
by others.14,21 On the other hand, an apparently successful
attempt was described by Kobayashi in 1976.21 According
to this, exposure of 20,30,50-tri-O-acetyl-8-bromoadenosine
(1g) to potassium fluoride in the presence of 18-crown-6 at
high temperature (120 ꢀC) for extended period of time
(48 h) led to the formation of 20,30,50-tri-O-acetyl-8-fluoro-
adenosine (1h) in moderate yield of 25%. However, no
attempts at removal of the acetyl protecting groups were
described in this communication. The identity of 20,30,50-
tri-O-acetyl-8-fluoroadenosine (1h) was confirmed in 1996
by Barrio, who was able to access the same compound
by fluorination of 20,30,50-tri-O-acetyl-adenosine (1i) with
X
RO
N
O
NH2
N
N
N
RO
OR
1a, R=H, X = H
1b, R=H, X = F
1c, R=H, X = Cl
1d, R=H, X = Br
1e, R=H, X = I
1f, R=Ac, X = NH2
1g, R=Ac, X = Br
1h, R=Ac, X = F
1i , R=Ac, X = H
1j , R=H, , X = N3
Figure 1.
* Corresponding author. Tel.: +1 732 594 8563; fax: +1 732 594 9556;
e-mail: gabor_butora@merck.com
0040–4020/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2007.02.053

G. Butora et al. / Tetrahedron 63 (2007) 3782–3789
3783
elemental fluorine.22 However, it was not until 1998 Barrio
described the first successful synthesis of 8-fluoroadenosine
(1b) by an enzymatic deprotection of 20,30,50-tri-O-acetyl-8-
fluoroadenosine (1h) using thermally stable hydrolases.14
Br
Br
HO
N
HO
N
O
NH2
O
N
NH2
N
i
N
N
N
N
O
O
HO
OH
Our interest in 8-fluoroadenosines stemmed from the
obvious capacity of the fluorine to attenuate the basicity of
the nitrogen atom at position 7. It is well known that this
nitrogen plays a crucial role in adenosine deaminase
(ADA) mediated metabolic degradation of adenosine.23
During the catalytic cycle an acid–base interaction between
7-N of adenosine and aspartate-296 of ADA increases the
electrophilicity of adenosine’s 6-C, which in turn leads to
ADA-catalyzed nucleophilic hydroxylation at this position
and eventually to the formation of inosine.24 Therefore,
8-fluoroadenosine (1b) should be a metabolically more
robust isostere of adenosine.
2
1d
iii
ii
NH2
N
N
N
F
O
HO
N
N
O
NH2
N
O
N
N
O
O
O
O
3
4
Scheme 1. Reagents and conditions: (i) 2,2-dimethoxypropane, PTSA,
DMF; (ii) CsF, MeCN, 100 ꢀC, sealed tube, 18 h; (iii) Ref. 27.
The chief obstacle in the successful synthesis of 8-fluoro-
adenosine (1b) is the removal of protecting groups in the
presence of a quite reactive C8–F bond. This was clearly
demonstrated by Barrio, who noticed a rapid fluorine
displacement in the presence of alkali, alkoxides, or ammo-
nia and was successful in removing the acetate protecting
groups only under the rather mild conditions of an enzymatic
hydrolysis.14 On the other hand, the decreased basicity of
7-N in 8-fluoroadenosine (1b) should make its protonation
more difficult, and, therefore, it should be comparatively
stable under acidic conditions. Clearly, the use of acid-labile
protecting groups appeared to be a logical choice.
The ease by which this cyclization could be monitored
(HPLC condition A) allowed for a facile optimization of
the halogen-exchange conditions. It soon became obvious
that the use of crown-ethers is not necessary for the exchange
and either KF or CsF is an effective fluoride source. The re-
action was conveniently performed at 100 ꢀC; it became
sluggish at lower temperature, and the yields diminished
when the reaction temperature was increased. The careful
exclusion of moisture was of crucial importance, and the
fluoride salt was flame dried under high vacuum before use.
In order to prevent the ring closure after the halogen
exchange, the 50-hydroxyl was protected as an acetate 5,
Scheme 2. Exposure of this material in acetonitrile to cesium
fluoride at 100 ꢀC for 9 h led to the formation of 50-O-acetyl-
8-fluoro-20,30-O-(1-methylethylidene) adenosine (6) (about
2. Chemistry
Even though direct C8-fluorination22 of a suitably protected
adenosine was certainly possible, we decided to adopt the
more facile halogen-exchange protocol.21,25 The starting
material, 8-bromoadenosine (1d) could be easily synthe-
sized on large scale by a direct bromination of adenosine
(1a) under controlled pH of 4.0.15 Initial attempts at dis-
placement of the bromine atom with fluorine in unprotected
1d using cesium or potassium fluoride led to complete con-
sumption of the starting material without production of any
detectable amounts (HPLC condition A or B) of the desired
8-fluoroadenosine (1b). In the absence of the fluorine source
1d was completely stable under otherwise identical
conditions.
O
Br
Br
HO
O
N
N
O
O
NH2
NH2
N
N
i
N
N
N
N
O
O
O
O
2
5
O
In order to improve the solubility of 8-bromoadenosine in
acetonitrile it was converted under standard conditions to
the known26 8-bromo-20,30-O-(1-methylethylidene) adeno-
sine (2). When this material was exposed to cesium or potas-
sium fluoride in dry acetonitrile at elevated temperature of
100 ꢀC, it smoothly converted to the 50,8-anhydro-adenosine
derivative 4, Scheme 1. Once again, under otherwise iden-
tical conditions, when the fluorine source was omitted, no
cyclization took place and the starting material was recov-
ered unchanged. This compound was previously synthesized
from bromo derivative 2 under strongly basic conditions.27
The stability of 2 in acetonitrile at 100 ꢀC in the absence
of the fluorine source suggested that the halogen exchange
indeed took place (intermediate 3) and this was followed
by a secondary ring closure.
ii
O
N
O
X
O
O
N
O
N
O
F
NH
N
O
N
N
O
O
N
O
NH2
N
NH2
5 or 6
N
N
N
N
O
O
O
O
7, X=Br
8, X=F
6
ii
Scheme 2. Reagents and conditions: (i) Ac2O, pyridine, 60 ꢀC, 18 h; (ii)
CsF, MeCN, 100 ꢀC, sealed tube, 9 h.

3784
G. Butora et al. / Tetrahedron 63 (2007) 3782–3789
50% conversion) in a 24% isolated yield. Extending the
reaction time to 18 h led to full consumption of the starting
material, unfortunately the yield of 6 rapidly decreased.
Careful LCMS analysis of the reaction mixtures indicated
formation of dimers (7, [M+H]+¼775.3 and 8, [M+H]+¼
715.3) resulting from reaction of the formed 50-O-acetyl-
8-fluoro-20,30-O-(1-methylethylidene) adenosine (6) with
unreacted 5 (compound 7) followed by an additional halogen
exchange (compound 8).
but more nucleophilic formic or acetic acid led to the for-
mation of acetate intermediate 13 and formate 14. Quite sur-
prisingly, the nucleophilic displacement of the fluorine did
not take place until both protecting groups were removed.
At later stages these reaction mixtures contained only the
8-fluoro derivative 3, ester 13 or 14, the desired 8-fluoro-
adenosine 1b, and the final product of hydrolysis, 8-oxo-
adenosine 15. Even though isolation of the desired
8-fluoroadenosine (1b) was possible, the yield was rather
disappointing. Monitoring of these hydrolyses was further
complicated by the fact that under most HPLC conditions,
the desired 8-fluoroadenosine (1b) and the 8-oxoadenosine
(15) were undistinguishable and conclusions were initially
drawn from MS data only. Later it was discovered that
omission of the modifier (TFA, formic acid, or ammonium
formate) from the mobile phase resulted in excellent peak
separation and this observation was applicable for prepara-
tive separations as well.
Since an acid-induced removal of the acetate in 6 was poten-
tially difficult, this protecting group was replaced by the
acid-labile tetrahydropyranyl, Scheme 3. Exposure of the
bromo derivative 9 to cesium fluoride in acetonitrile at
100 ꢀC for 12 h resulted in a clean halogen exchange and
the 8-fluoro derivative 10 could be isolated in 28% yield as
a mixture of THP-derived diastereoisomers. Extension of
the reaction time, and/or elevation of the temperature led
to full consumption of the starting material, but increased
the formation of dimers 11 ([M+H]+¼859.6) and 12
([M+H]+¼799.7) and considerably lowered the yield. Even-
tually, the formation of these nucleoside oligomers could be
greatly suppressed by performing the reaction under high
dilution conditions.
F
N
F
HO
N
O
N
O
O
NH2
i
O
iii
NH2
N
N
N
N
N
O
O
O
O
3
10
Br
Br
HO
N
ii
O
N
O
O
NH2
N
O
NH2
N
i
N
XO
F
N
N
N
O
O
HO
HO
N
N
O
O
O
O
NH2
NH2
N
N
2
N
N
9
N
N
HO
OH
HO
OH
O
13, X = CH3CO
14, X = HCO
1b
O
ii
O
X
O
ii, iii or iv
O
iii
N
O
N
N
F
HO
NH
NH
N
O
N
N
O
O
O
NH2
N
N
O
NH2
O
N
O
NH2
9 or 10
N
N
N
N
HO
OH
N
N
O
O
O
O
15
Scheme 4. Reagents and conditions: (i) 1% TFA, water, 45 min; (ii)
10% aqueous perchloric acid, 1 h; (iii) aqueous acetic or formic acid,
20–40 ꢀC; (iv) 20–80% TFA, water.
11, X=Br
12, X=F
10
ii
Scheme 3. Reagents and conditions: (i) 3,4-dihydro-2H-pyran, camphor
sulfonic acid, MeCN, DMF, 0 ꢀC, 2 h; (ii) CsF, MeCN, 100 ꢀC, 12 h.
Finally, the clean removal of the acetonide from the fluoro
derivative 3 was accomplished by the use of perchloric
acid, an extremely weak nucleophile. Exposure of 3 to
a 10% aqueous solution of perchloric acid resulted in full
cleavage of the acetonide in approximately 1 h at which
time only traces of 15 were detectable. The product was
then easily purified by preparative HPLC (condition D)
and solid samples were obtained by lyophilization. No
appreciable decomposition was noticeable during these
operations and significant quantities of 8-fluoroadenosine
(1b) were synthesized.
Exposure of the protected fluoro derivative 10 to even
weakly acidic medium, such as 1% TFA, led to rapid loss
of the THP protecting group, Scheme 4. Under these con-
ditions the acetonide remained unchanged, and the fluoro
derivative 3 could be easily intercepted. Surprisingly, the
fluoride in 3 was found to be perfectly stable, and no
8-oxo analogs were detectable (LCMS). The removal of
the acetonide protecting group, however, required somewhat
more forceful conditions. Increasing the concentration of tri-
fluoroacetic acid led to rapid removal of both the THP and
the acetonide, however, considerable amounts of 8-oxo-
adenosine (15), the chief breakdown product of 8-fluoro-
adenosine (1b), were formed, Scheme 4. The use of weaker
It is reasonable to anticipate that a temporary introduction of
the fluorine, as depicted in Scheme 1, would mediate not
only an intramolecular cyclization, but also in fact it could

G. Butora et al. / Tetrahedron 63 (2007) 3782–3789
3785
Table 1. Proton and carbon chemical shifts of 8-fluoroadenosine (CD3OD,
24 ꢀC)
facilitate any nucleophilic displacement and examples of
such acceleration have been reported.25 When 8-bromo-
adenosine (1d) is reacted with sodium azide (80 ꢀC, DMF)
the bromide is readily displaced to yield 8-azidoadenosine
(1j).5 However, any attempt to produce 8-cyanoadenosine
using this simple procedure fails. In fact, 8-cyanoadenosine
is accessible from the corresponding iodide and zinc cya-
nides only by a tetrakis[tri(2-furyl)phosphine]palladium(0)
catalyzed cross-coupling.28 According to expectations,
when the protected 8-bromoadenosine 9 was heated with
sodium cyanide in acetonitrile at 100 ꢀC, the starting
material remained unchanged. However, addition of cesium
fluoride resulted in smooth conversion of the starting mate-
rial to 8-cyanoadenosine derivative 16 and we attribute this
acceleration to an intermediate formation of the 8-fluoro
derivative 10, Scheme 5. The scope and limitations of such
reactions for preparation of 8-substituted adenosine deriva-
tives are currently being examined.
8-Fluoroadenosine
1H d ppm, (Hz)
13C d ppm (Hz)
10
20
300
4
5.84 (s, 6.9)
4.91 (t, 6.2)
89.74
64.15
72.74
88.67
4.33 (dd, 5.0, 1.8)
4.14 (br d, 2.3)
3.84 (dd, 12.6, 2.5)
3.70 (dd, 12.6, 3.0)
8.16 (s)
500
50
2
63.80
153.06
148.85
4
5
6
8
115.28
156.72
152.45 (d, 250.9)
For details see Section 5 as well as Supplementary data.
Table 2
Sample
pKa
Br
NC
O
N
O
N
O
O
Experimental
(ISA H2O at 25 ꢀC)
Calculated
(ACD/Labs 8.00)
O
O
NH2
NH2
N
N
i
X
N
N
N
N
1a
1b
3.61ꢂ0.01
3.18ꢂ0.01
3.40
2.72
O
O
O
O
9
16
F
Table 3
O
N
O
O
NH2
N
pH
Composition
t1/2 (days)
ii
N
N
2.0
3.0
5.0
7.0
9.0
9.5
10.0
10.5
11.0
0.1 M Perchloric acid/1 M NaOH
0.1 M Perchloric acid/1 M NaOH
0.1 M Perchloric acid/1 M NaOH
0.1 M Perchloric acid/1 M NaOH
0.1 M NaOH/1 M HCl
0.1 M NaOH/1 M HCl
0.1 M NaOH/1 M HCl
0.1 M NaOH/1 M HCl
0.1 M NaOH/1 M HCl
9.7
22.6
186.3
>250a
13.8
O
O
10
Scheme 5. Reagents and conditions: (i) NaCN, acetonitrile, 100 ꢀC, 12 h;
7.0
1.4
0.12
0.029
(ii) NaCN, CsF, acetonitrile, 100 ꢀC, 12 h.
a
3. Properties and stability of 8-fluoroadenosine
Estimated value. No significant hydrolysis noticed at this pH.
The significant quantities of 8-fluoroadenosine (1b) avail-
able through the above described procedure enabled us to
examine its physicochemical properties in closer detail.
The pH-dependent hydrolytic stability of 1b was examined
in close detail and the results are summarized in Table 3.
An estimated half-life value in excess of 250 days indicates
that 8-fluoroadenosine is remarkably stable at neutral pH. In
agreement with its predicted behavior, 1b is considerably
more stable under acidic conditions (t1/2¼9.7 days, pH¼
2.0) than under basic conditions (t1/2¼42 min, pH¼11.0).
The half-life values suggest that under physiological condi-
tions 8-fluoroadenosine might be in fact hydrolytically quite
stable.
Its spectral characteristics, including 1H, 13C, and 19F NMR
spectra were in perfect agreement with the structure 8-
fluoroadenosine, and even though some of the signals were
previously assigned incorrectly,14 the reported values closely
matched those recorded by us. Their correct assignment is
summarized in Table 1.
The UV spectrophotometric profile of 1b (lmax¼252.0 nm)
showed a small hypsochromic shift in comparison with
adenosine (lmax¼260.0 nm). Also the optical rotation of
1b [a]D –54.7 (c 0.92 g/100 mL, water) was close to that
recorded under identical conditions for adenosine (ꢁ59.7).
4. Conclusion
A facile non-enzymatic synthesis of 8-fluoroadenosine (1b)
was developed through which significant quantities of this
simple nucleoside became available for the first time. This
allowed for a detailed examination of its physicochemical
properties, including basicity and pH-dependent stability.
The obtained data suggest that 8-fluoroadenosine, or its
derivatives, might be more stable at physiological conditions
than expected.
In agreement with calculated values, the basicity of 1b
(pKa¼3.18) is lower than that of adenosine (1a, pKa¼
3.61), Table 2. According to this, substantial protonation
(>50%) can be expected at pH values less than 3.2, and
this in turn should translate to markedly increased rate of
hydrolysis only at very low pH values.

3786
G. Butora et al. / Tetrahedron 63 (2007) 3782–3789
5. Experimental
5.2. 8-Bromoadenosine (1d)15
5.1. General
An aqueous solution of sodium acetate (0.5 M, 200 mL) was
treated with glacial acetic acid until the pH reached 4.00 (pH
electrode, Accumet AP62, 13-620-AP50). To this buffer
adenosine (1a, 5.15 g, 19.3 mmol) was added and stirred
with gentle heating until the solid has fully dissolved. The
solution was allowed to cool to ambient temperature and
a mixture of bromine in water (4.60 g, 57.6 mmol in
100 mL of water) was added. Vigorous stirring was contin-
ued for 4 h, after which time no starting material was detect-
able by HPLC (condition B). The reaction mixture was
decolorized by the addition of solid sodium bisulfite (spatula
tip) and the pH was adjusted to 7.00 with 50% aqueous
solution of sodium hydroxide. The suspension was allowed
to stand at +5 ꢀC overnight, and the solid was collected by
filtration. The filter cake was washed with water (3ꢃ20 mL)
and acetone (1ꢃ20 mL), and dried on high vacuum to
yield 5.30 g (79%) of the pure product. tR(A)¼2.67 min,
All non-hydrolytic reactions, unless indicated otherwise,
were carried out in dry solvents purchased from Aldrich.
NMR spectra were recorded on a Varian Inova 500 or
600 MHz spectrometer and the shifts are positive in the
low field direction. The 19F chemical shifts were referenced
indirectly to CFCl3 using the deuterium signal of the solvent
(CD3OD). Elemental analyses were performed by Robertson
Microlit Laboratories, Madison, NJ.
Analytical HPLC condition A: Agilent 1100, Waters Atlan-
tis dC18, 5 mm, 4.6ꢃ50 mm at 37 ꢀC; eluant: water and
MeCN without a modifier; gradient (% organic): 0.0 min
(0), 5.0 min (100), 6.0 min (100); detector: PDA UV Detec-
tor/Waters Micromass ZQ Mass Spectrometer. Analytical
HPLC condition B: identical to A except column: Waters
SunFire C18, 5 mm, 4.6ꢃ100 mm. Preparative HPLC con-
dition C: Waters 2525 pump, 2767 injector/collector 2996
Waters Sunfire (50ꢃ100 mm, 5 mm); eluant: water and
MeCN without a modifier; gradient (% organic): 0.0 min
(25), 1.5 min (25), 10.0 min (65), 50 mL/min; detector:
PDA UV Detector/Micromass ZQ Single Quad Mass
Spectrometer, mass-triggered collection. Preparative HPLC
condition D: same as condition C except gradient (% or-
ganic): 0.0 min (3), 2.0 min (3), 10.5 min (15), 10.7 min
(100), 13.7 min (100).
1
tR(B)¼2.73 min. H NMR (500 MHz, DMSO-d6), d: 8.12
(s, 1H), 7.60 (br s, 2H), 5.83 (d, J¼6.9, 1H), 5.46 (br s,
2H), 5.24 (br s, 2H), 5.08 (t, J¼6.0, 1H), 4.20 (dd, J¼5.0,
2.0 Hz, 1H), 3.98 (dd, J¼6.2, 3.7 Hz, 1H), 3.68 (dd,
J¼12.1, 3.9 Hz, 1H), 3.52 (dd, J¼12.1, 4.1 Hz, 1H). 13C
NMR (125 MHz, DMSO-d6), d: 155.1, 152.3, 149.9,
127.2, 119.7, 90.4, 86.7, 71.1, 70.9, 62.1. HRMS for
C10H12BrN5O4 calculated [M+H]+: 346.0151, found:
346.0141. For C10H12BrN5O4 calculated: 34.70% C,
3.49% H; found: 34.48% C, 3.35% H.
The optical rotation was determined at 20 ꢀC and 589 nm on
a Perkin–Elmer 341 polarimeter, using a 1 dm cell.
5.3. 8-Bromo-20,30-O-(1-methylethylidene) adenosine (2)
A solution of 8-bromoadenosine (1d, 5.28 g, 15.3 mmol),
para-toluenesulfonic acid (monohydrate, 3.24 g, 16.8 mmol),
and 3,3-dimethoxypropane (30 mL) in dimethylformamide
(5 mL) was stirred at ambient temperature for 30 h. The
solvent was removed in vacuo and the residue was purified
by gradient chromatography [silica gel, 2-propanol/chloro-
form (2-propanol: 0–10%)] to yield 3.60 g (61%) of
the pure product. tR(A)¼3.63 min, tR(B)¼3.94 min. 1H NMR
(500 MHz, DMSO-d6), d: 8.14 (s, 1H), 6.01 (d, J¼2.8, 1H),
5.59 (m, 1H), 5.02 (dd, J¼6.2, 2.8 Hz, 1H), 4.16 (m, 1H),
3.51 (dd, J¼11.7, 5.7 Hz, 1H), 3.43 (dd, J¼11.5, 5.7 Hz,
1H), 1.54 (s, 3H), 1.32 (s, 3H). 13C NMR (125 MHz,
DMSO-d6), d: 155.0, 152.8, 149.7, 126.4, 119.3, 113.2,
91.0, 87.1, 81.9, 81.6, 61.5, 27.1, 25.2. HRMS for
C13H16BrN5O4 calculated [M+H]+: 386.0464, found:
368.0466. For C13H16BrN5O4 calculated: 40.43% C, 4.18%
H; found 40.42% C, 4.16% H.
pH-dependent stability: a stock solution of 1b (1 mg/mL in
MeCN) was prepared. An aliquot (0.1 mL) was added to
a 4 mL vial and blown dry with N2 and mild heat. Once
dried, 1 mL of buffer (see Table 2) was added and gently
shaken to mix. The solution was then transferred to an
HPLC vial and analyzed at time by HPLC/DAD/MS analysis
under the following conditions: isocratic [A¼water, B¼
acetonitrile, 5% B for 10 min, 1 mL/min] on the column
Sunfire C18 (4.6ꢃ100 mm, 5 mm) at 40 ꢀC with diode array
detection at 252 nm and injection volume of 10 mL. Each
sample was prepared in duplicate and equilibrated at room
temperature (w22 ꢀC).
The pKa values were determined by potentiometric titrations
with spectrophotometric analysis. The titrations were
performed with the Sirius GLpKa/D-PAS using a double
junction electrode. The electrode was standardized from
pH 1.8 to 12.2. The KOH titrant was standardized against
potassium hydrogen phthalate and was approximately
0.5 M. The sample was dissolved in ionic strength adjusted
(ISA) H2O (0.15 M KCl) to create a stock solution of
4.0 mg/mL. An aliquot of the stock solution (0.075 mL for
adenosine and 0.05 mL for L-001855881) was added to
the titration vial and diluted up to 15 mL of ISA H2O. The
starting pH of the solution was adjusted with 0.5 M HCl.
The solution was titrated from pH 1.8 to 7.0 at 25 ꢀC in
ISA H2O. Titrations were repeated several times. UV spectra
were recorded from 200 to 700 nm and the data were ana-
lyzed from 250 to 360 nm. The pKa values were calculated
using Refinement Pro v.1.0 software.
5.4. (3aR,4R,13R,13aR)-2,2-Dimethyl-4,5,13,13a-tetra-
hydro-3aH-4,13-epoxy[1,3]dioxolo[5,6][1,3]oxazo-
cino[3,2-e]purin-8-amine (4)
A solution of 8-bromo-20,30-O-(1-methylethylidene) adeno-
sine (2, 200 mg, 0.52 mmol) and cesium fluoride (790 mg,
5.2 mmol) in dry acetonitrile (60 mL) was heated to
100 ꢀC in a closed thick walled glass reaction vessel for
18 h. The reaction mixture was allowed to cool to ambient
temperature, the solid was filtered off, and the solvent was
evaporated in vacuo. The residue (198 mg) was purified by
preparative TLC (CHCl3/iPrOH 9:1) to yield 109.4 mg
(69%) of the pure product. tR(A)¼3.32 min, tR(B)¼3.56 min.

G. Butora et al. / Tetrahedron 63 (2007) 3782–3789
3787
1H NMR (500 MHz, DMSO-d6), d: 8.11 (s, 1H), 7.10 (s,
1H), 5.94 (s, 1H), 5.10 (d, J¼5.5 Hz, 1H), 4.90 (d,
J¼5.7 Hz, 1H), 4.74 (s, 1H), 4.64 (dd, J¼13.0, 1.6 Hz,
1H), 4.13 (d, J¼13.1 Hz, 1H), 1.46 (s, 3H), 1.29 (s, 3H).
13C NMR (500 MHz, DMSO-d6), d: 154.7, 153.1, 152.0,
147.6, 114.4, 111.9, 85.8, 85.2, 84.8, 80.1, 74.3, 25.9,
24.3. HRMS for C13H15FN5O4 calculated [M+H]+:
306.1202, found: 306.1214. For C13H15N5O4 calculated:
51.14% C, 4.95% H; found: 51.08% C, 5.05% H.
0 ꢀC was continued for 2 h, and the reaction was quenched
with 10% NaHCO3 (40 mL). The product was extracted
with chloroform (2ꢃ250 mL), the combined extracts were
dried (anhydrous sodium sulfate), and concentrated in
vacuo. The residue was purified by gradient chromatography
[silica gel, ethyl acetate/hexanes (ethyl acetate: 0–100%)]
to yield 3.83 g (66%) of the product as a mixture of dia-
stereoisomers. tR(A)¼4.54 min, tR(B)¼5.02 min. 1H NMR
(500 MHz, DMSO-d6): major isomer, d: 8.14 (m, 1H),
6.05 (m, 1H), 5.76 (m, 2H), 4.48 (m, 1H), 4.26 (m, 1H), 3.68
(m, 1H), 3.64 (m, 2H), 3.36 (m, 1H), 3.29 (m, 1H), 1.61 (m,
2H), 1.52 (m, 3H), 1.37 (m, 4H), 1.32 (m, 3H); minor isomer,
d: 8.14 (m, 1H), 6.05 (m, 1H), 5.08 (m, 2H), 4.37 (m, 1H),
4.26 (m, 1H), 3.58 (m, 1H), 3.48 (m, 2H), 3.31 (m, 1H),
3.18 (m, 1H), 1.52 (m, 3H), 1.49 (m, 2H), 1.37 (m, 4H),
1.32 (m, 3H). 13C NMR (500 MHz, DMSO-d6), major iso-
mer, d: 155.0, 152.9, 149.9, 128.9, 126.5, 119.1, 113.3,
98.0, 90.5, 86.0, 82.3, 81.7, 66.5, 61.1, 29.9, 27.0, 25.3,
18.8; minor isomer, d: 155.0, 152.9, 149.8, 128.1, 126.5,
119.1, 113.2, 97.7, 90.5, 85.6, 82.1, 81.6, 66.3, 60.8, 29.8,
27.0, 24.9, 18.7. HRMS for C18H24BrN5O5H calculated:
470.1039, found: 470.1036. For C18H24BrN5O4 (CH3CN)
calculated: 46.97% C, 5.32% H; found: 47.40% C, 5.08% H.
5.5. 50-O-Acetyl-8-bromo-20,30-O-(1-methylethylidene)
adenosine (5)
A solution of 8-bromo-20,30-O-(1-methylethylidene) adeno-
sine (2, 3.00 g, 7.77 mmol) in anhydrous pyridine (40 mL)
was treated with acetic anhydride (1.10 mL, 11.0 mmol)
and the reaction mixture was stirred at 60 ꢀC overnight.
The solvent was removed in vacuo and the residue was puri-
fied by gradient chromatography [silica gel, 2-propanol/
chloroform (iPrOH: 0–10%)] to afford 3.15 g (95%) of the
pure material. tR(A)¼4.10 min, tR(B)¼4.48 min. 1H NMR
(500 MHz, DMSO-d6), d: 8.50 (s, 1H), 7.52 (br s, 1H), 6.07
(s, 1H), 5.70 (d, J¼6.2 Hz, 1H), 5.12 (dd, J¼6.0, 3.7 Hz,
1H), 4.31 (m, 1H), 4.22 (dd, J¼11.7, 5.0 Hz, 1H), 4.06
(dd, J¼11.7, 7.3 Hz, 1H), 1.93 (s, 3H), 1.52 (s, 3H), 1.30
(s, 3H). 13C NMR (500 MHz, DMSO-d6), d: 170.0, 154.9,
152.9, 149.8, 126.3, 119.2, 113.5, 90.3, 84.5, 82.3, 81.3,
63.3, 27.0, 25.2, 20.4. HRMS for C15H18BrN5O5 calculated
[M+H]+: 428.0569, found: 428.0562. For C15H18BrN5O4
calculated: 42.07% C, 4.24% H; found: 41.96% C, 4.33% H.
5.8. 8-Fluoro-20,30-O-(1-methylethylidene)-50-O-(tetra-
hydro-2H-pyran-2-yl) adenosine (10)
A thick walled glass pressure reaction vessel, equipped with
a magnetic stirring bar was charged with cesium fluoride
(650 mg, 4.30 mmol), closed under a septum, and flame dried
at high vacuum. The vessel was flooded with nitrogen and
allowed to cool to ambient temperature. A solution of
8-bromo-20,30-O-(1-methylethylidene)-50-O-(tetrahydro-2H-
pyran-2-yl) adenosine (9, 200 mg, 0.43 mmol) in acetonitrile
(60 mL) was added via syringe, the vessel was closed with
a solid Teflon bushing, and stirred at 100 ꢀC. After 12 h
when the HPLC analysis indicated a conversion of about
50%, the reaction mixture was allowed to cool to room tem-
perature. The solid was filtered off, and the solvent was
removed in vacuo. The crude product was purified by prepar-
ative HPLC (condition C) to yield 48.0 mg (28%) of the pure
product and 12.0 mg (6%) of unreacted starting material.
tR(A)¼4.11 and 4.36 min, tR(B)¼4.52 and 4.84 min. Faster
5.6. 50-O-Acetyl-8-fluoro-20,30-O-(1-methylethylidene)
adenosine (6)
A solution of 50-O-acetyl-8-bromo-20,30-O-(1-methylethyl-
idene) adenosine (5, 200 mg, 0.47 mmol) and cesium
fluoride (710 mg, 4.70 mmol) in dry acetonitrile (60 mL)
was heated to 100 ꢀC in a closed thick walled glass reaction
vessel, until HPLC analysis (condition B) indicated about
50% conversion, for about 9 h. The solid was filtered off
and the filtrate was concentrated in vacuo (210 mg). The
crude product was purified by preparative HPLC (condition
C) to afford 41.6 mg (24%) of the pure product alongside
with 36.7 mg of recovered starting material. tR(A)¼3.94 min,
1
eluting isomer: H NMR (500 MHz, CDCl3), d: 8.30 (br s,
1
tR(B)¼4.30 min. H NMR (500 MHz, CD3OD), d: 8.36 (s,
1H), 6.06 (s, 1H), 5.90 (br s, 2H), 5.60 (d, J¼6.0 Hz, 1H),
5.08 (m, 1H), 4.54 (m, 1H), 4.40 (m, 1H), 3.50 (m, 1H), 3.82
(dd, J¼11.0, 5.7 Hz, 1H), 3.70 (m, 1H), 3.54 (dd, J¼11.0,
4.7 Hz, 1H), 3.38 (m, 1H), 1.60 (s, 3H), 1.40 (s, 3H), 1.8–1.4
(br m,w5H). 13C NMR (125 MHz, CDCl3), d: 154.2, 152.2,
152.1, 150.1, 148.3, 114.4, 98.9, 88.6, 85.9, 82.8, 81.5, 66.8,
62.0, 30.1, 27.2, 25.4, 25.2, 19.2. 19F NMR (CD3OD), d:
ꢁ105.94. Slow eluting isomer: 1H NMR (500 MHz,
CDCl3), d: 8.30 (br s, 1H), 6.06 (d, J¼1.6 Hz, 1H), 5.90 (br
s, 2H), 5.56 (d, J¼6.2 Hz, 1H), 5.06 (m, 1H), 4.52 (m, 1H),
4.40 (m, 1H), (dd, J¼10.8, 5.0 Hz, 1H), 3.90 (m, 1H), 3.76
(m, 1H), 3.54 (dd, J¼10.7, 7.1 Hz, 1H), 3.42 (m, 1H), 1.60
(s, 3H), 1.40 (s, 3H), 1.8–1.4 (br m, w5H). 13C NMR
(125 MHz, CDCl3), d: 154.3, 152.2, 152.0, 150.0, 148.1,
114.5, 98.9, 88.3, 86.4, 83.1, 81.7, 67.2, 62.0, 30.2, 27.2,
25.5, 25.2, 19.1. 19F NMR (CD3OD), d: ꢁ105.47. HRMS
(isomers not resolved) for C18H24FN5O5 calculated:
410.1837, found: 410.1840. For C18H24FN5O5 (trihydrate)
calculated: 46.65% C, 6.52% H; found: 46.74% C, 6.26% H.
1H), 6.13 (d, J¼1.8 Hz, 1H), 5.60 (dd, J¼6.2, 1.6 Hz, 1H),
5.09 (dd, J¼6.2, 3.4 Hz, 1H), 4.41 (m, 1H), 4.20 (dd, J¼8.1,
6.9 Hz, 2H), 1.96 (s, 3H), 1.57 (s, 3H), 1.38 (s, 3H). 13C
NMR (500 MHz, CD3OD), d: 172.3, 153.9, 153.1, 151.8,
148.6, 115.7, 90.2, 86.9, 84.6, 82.7, 64.6, 27.3, 25.4, 20.5.
19F NMR (CD3OD) d ꢁ104.0. HRMS for C15H18FN5O5
calculated [M+H]+: 368.1370, found: 368.1383. For
C15H18FN5O5 (dihydrate) calculated: 44.66% C, 5.50% H;
found: 44.24% C, 4.41% H.
5.7. 8-Bromo-20,30-O-(1-methylethylidene)-50-O-(tetra-
hydro-2H-pyran-2-yl) adenosine (9)
A solution of 8-bromo-20,30-O-(1-methylethylidene) adeno-
sine (2, 4.76 g, 12.3 mmol), camphor sulfonic acid (6.3 g,
27.1 mmol) in a mixture of acetonitrile (50 mL) and
dimethylformamide (5 mL) was treated at 0 ꢀC with
3,4-dihydro-2H-pyran (11.2 mL, 123 mmol). Stirring at

3788
G. Butora et al. / Tetrahedron 63 (2007) 3782–3789
5.9. 8-Fluoro-20,30-O-(1-methylethylidene) adenosine (3)
solution of 8-fluoro-20,30-O-(1-methylethylidene)-50-
3.64 (dd, J¼11.0, 4.6 Hz, 1H), 3.60 (dd, J¼8.5, 3.0 Hz,
1H), 3.3 (m, 1H), 1.60 (s, 3H), 1.2–1.6 (br m, 6H), 1.40 (s,
3H). 13C NMR (125 MHz, CD3OD), d: 158.3, 156.7,
150.1, 124.9, 115.3, 112.1, 99.9, 93.3, 88.3, 85.2, 83.4,
68.0, 63.0, 31.2, 27.4, 26.4, 25.5, 20.1. For C19H24N6O5 cal-
culated [M+H]+: 416.18, found: 417.23. Slow eluting
isomer: 20.1 mg (12%). tR(A)¼4.46 min, tR(B)¼4.92 min.
1H NMR (500 MHz, CD3OD), d: 8.28 (s, 1H), 6.24 (d,
J¼5.79 Hz), 5.69 (dd, J¼6.4, 1.8 Hz, 1H), 5.09 (dd,
J¼6.2, 3.2 Hz, 1H), 4.44 (m, 2H), 3.84 (dd, J¼11.0,
5.3 Hz, 1H), 3.73 (ddd, J¼8.5, 8.2, 2.7 Hz, 1H), 3.51 (dd,
J¼11.0, 6.6 Hz, 1H), 3.38 (m, 1H), 3.30 (m, 1H), 1.59 (s,
3H), 1.39 (s, 3H), 1.32–1.62 (br m, 6H). 13C NMR
(125 MHz, CD3OD), d: 15.8.4, 156.8, 150.1, 125.0, 121.1,
116.4, 115.4, 111.9, 100.4, 92.6, 88.3, 84.8, 83.4,
68.2, 63.2, 31.3, 27.4, 26.4, 25.6, 20.1. For C19H24N6O5
calculated [M+H]+: 416.18, found: 417.20. HRMS
(isomers not resolved) for C19H24N6O5 calculated
[M+H]+: 417.1886, found: 417.1896. For C19H24N6O5
(hemihydrate) calculated: 53.54% C, 5.92% H; found:
53.14% C, 5.71% H.
A
O-(tetrahydro-2H-pyran-2-yl) adenosine (10, 30 mg,
0.073 mmol) in acetonitrile (2 mL) was treated with five
drops of trifluoroacetic acid and stirred at ambient tempera-
ture for 45 min. The entire reaction mixture was injected
onto a preparative HPLC and purified under condition D.
The fractions containing the product were pooled and lyoph-
ilized to afford 14.9 mg (63%) of the desired product.
tR(A)¼3.43 min, tR(B)¼3.71 min. 1H NMR (500 MHz,
CD3OD), d: 8.16 (s, 1H), 7.13 (s, 2H), 5.96 (s, 1H), 5.58
(d, J¼5.4 Hz, 1H), 4.96 (d, J¼3.4 Hz, 1H), 4.14 (br s, 1H),
1.52 (s, 3H), 1.32 (s, 3H). 13C NMR (125 MHz, CD3OD), d:
155.2, 152.6, 151.0, 149.3, 147.6, 113.3, 88.2, 87.4, 82.2,
81.3, 61.3, 27.0, 25.2. 19F NMR (CD3OD), d ꢁ104.95.
HRMS for C13H16FN5O4 calculated [M+H]+: 326.1266,
found: 306.1245. For C13H16FN5O4 (hemihydrate) calcu-
lated: 46.71% C, 5.13% H; found: 46.69% C, 4.84% H.
5.10. 8-Fluoroadenosine (1b)
A
solution of 8-fluoro-20,30-O-(1-methylethylidene)-50-
O-(tetrahydro-2H-pyran-2-yl) adenosine (10, 43 mg,
0.105 mmol) was dissolved in 10% aqueous perchloric
acid and stirred at ambient temperature. The reaction prog-
ress was monitored by HPLC. The reaction was complete
after approximately 1 h, and 8-oxo-adenosine started to
appear. The reaction mixture was terminated by injection
into a preparative HPLC (condition D). The fractions con-
taining the product were pooled, and the solvent was
removed by lyophilization. In this fashion, 12.4 mg (46%)
Supplementary data
Supplementary data associated with this article can be found
in the online version, at doi:10.1016/j.tet.2007.02.053.
References and notes
of clean product was obtained. tR(A)¼2.48 min, tR(B)
¼
1. Sekine, M.; Okada, K.; Seio, K.; Kakeya, H.; Osada, H.; Sasaki,
T. Bioorg. Med. Chem. 2004, 12, 5193–5201.
2. Van Aerschot, A. A.; Mamos, P.; Weyns, N. J.; Ikeda, S.; De
Clercq, E.; Herdewijn, P. A. J. Med. Chem. 1993, 36, 2938–
2942.
3. Holmes, R. E.; Robins, R. K. J. Am. Chem. Soc. 1964, 86,
1242–1245.
4. Chatgilialoglu, C.; Navacchia, M. L.; Postigo, A. Tetrahedron
Lett. 2006, 47, 711–714.
5. Frieden, M.; Avino, A.; Eritja, R. Nucleosides Nucleotides
Nucleic Acids. 2003, 22, 193–202.
6. Kohyama, N.; Katashima, T.; Yamamoto, Y. Synthesis 2004,
17, 2799–2804.
7. Mamos, P.; Vanaerschot, A. A.; Weyns, N. J.; Herdewijn, P. A.
Tetrahedron Lett. 1992, 33, 2413–2416.
8. Brentnall, H. J.; Hutchinson, D. W. Tetrahedron Lett. 1972, 13,
2595–2596.
1
2.41 min. [a]D ꢁ54.7 (c 0.92 g/100 mL, water). H NMR
(500 MHz, CD3OD), d: 8.16 (s, 1H, 2-H), 5.84 (d,
J¼6.9 Hz, 1H, 10-H), 4.91 (t, J¼6.2 Hz, 1H, 20-H), 4.33
(dd, J¼5.0, 1.8 Hz, 1H, 30-H), 4.14 (br d, J¼2.3 Hz, 1H,
40-H), 3.84 (dd, J¼12.6, 2.5 Hz, 1H, 500-H), 3.70 (dd,
J¼12.6, 3.0 Hz, 1H, 50-H). 13C NMR (150 MHz, CD3OD),
d: 156.72 (6-C), 153.06 (2-C), 152.45 (d, J¼250.9 Hz, 8-
C), 148.85 (4-C), 115.28 (5-C), 89.74 (10-C), 88.67 (40-C),
64.15 (20-C), 72.74 (30-C), 63.80 (50-C). 19F NMR
(CD3OD), d ꢁ107.6. HRMS for C10H12FN5O4 calculated
[M+H]+: 286.0952, found: 286.0971. For C10H12FN5O4 cal-
culated: 44.15% C, 4.72% H, 23.40% N; found: 43.18% C,
4.11% H, 22.10% N.
5.11. 8-Cyano-20,30-O-(1-methylethylidene)-50-O-(tetra-
hydro-2H-pyran-2-yl) adenosine (16)
9. Ikehara, M.; Yamada, K. Chem. Pharm. Bull. (Tokyo) 1971, 19,
104–109.
10. Jansons, J.; Maurinsh, Y.; Lidaks, M. Nucleosides Nucleotides
1995, 14, 1709–1724.
11. Manfredini, S.; Baraldi, P. G.; Bazzanini, R.; Marangoni, M.;
Simoni, D.; Balzarini, J.; De Clercq, E. J. Med. Chem. 1995,
38, 199–203.
12. Lang, P.; Gerez, C.; Tritsch, D.; Fontecave, M.; Biellmann,
J. F.; Burger, A. Tetrahedron 2003, 59, 7315–7322.
13. Ratsep, P. C.; Robins, R. K.; Vaghefi, M. M. Nucleosides
Nucleotides 1990, 9, 197–204.
A flame-dried mixture of potassium cyanide (280 mg,
4.30 mmol) and cesium fluoride (196 mg, 1.29 mmol) was
treated with a solution of 8-bromo-20,30-O-(1-methylethyl-
idene)-50-O-(tetrahydro-2H-pyran-2-yl) adenosine (9, 200 mg,
0.43 mmol) in 100 mL of acetonitrile. The mixture was
heated in a thick walled pressure reaction vessel at 100 ꢀC
overnight. The mixture was allowed to cool to ambient tem-
perature, the solid was filtered off, and the residue was puri-
fied by preparative HPLC, condition C. Fast eluting isomer:
1
58.2 mg (33%). tR(A)¼4.43 min, tR(B)¼4.88 min. H NMR
(500 MHz, CD3OD), d: 8.28 (s, 1H), 6.24 (d, J¼2.1 Hz,
1H), 5.80 (dd, J¼6.2, 1.8 Hz, 1H), 5.07 (dd, J¼6.2,
2.7 Hz, 1H), 4.49 (m, 2H), 3.78 (dd, J¼10.8, 5.5 Hz, 1H),
14. Barrio, J. R.; Namavari, M.; Keen, R. E.; Satyamurthy, N.
Tetrahedron Lett. 1998, 39, 7231–7234.
15. Ikehara, M.; Kaneko, M. Tetrahedron 1970, 26, 4251–4259.

G. Butora et al. / Tetrahedron 63 (2007) 3782–3789
3789
16. Prakash, T. P.; Kumar, R. K.; Ganesh, K. N. Tetrahedron 1993,
49, 4035–4050.
17. Moriarty, R. M.; Vaid, R. K.; Hopkins, T. E.; Vaid, B. K.;
Prakash, O. Tetrahedron Lett. 1990, 31, 201–204.
18. Ikehara, M.; Ogiso, Y.; Maruyama, T. Chem. Pharm. Bull.
(Tokyo) 1977, 25, 575–578.
23. Cristalli, G.; Costanzi, S.; Lambertucci, C.; Lupidi, G.;
Vittori, S.; Volpini, R.; Camaioni, E. Med. Res. Rev. 2001,
21, 105–128.
24. Sideraki, V.; Mohamedali, K. A.; Wilson, D. K.; Chang, Z.;
Kellems, R. E.; Quiocho, F. A.; Rudolph, F. B. Biochemistry
1996, 35, 7862–7872.
19. Ryu, E. K.; MacCoss, M. J. Org. Chem. 1981, 46, 2819–2823.
20. Ikehara, M.; Yamada, S. J. Chem. Soc., Chem. Commun. 1968,
1509.
21. Kobayashi, Y.; Kumadaki, I.; Ohsawa, A.; Murakami, S.
J. Chem. Soc., Chem. Commun. 1976, 430–431.
22. Barrio, J. R.; Namavari, M.; Phelps, M. E.; Satyamurthy, N.
J. Am. Chem. Soc. 1996, 118, 10408–10411.
25. Subrayan, R. P.; Rasmussen, P. G. Tetrahedron 1995, 51, 6167–
6178.
26. Ikehara, M.; Tada, H.; Kaneko, M. Tetrahedron 1968, 24,
3489–3498.
27. Ikehara, M.; Kaneko, M.; Okano, R. Tetrahedron 1970, 26,
5675–5682.
28. Gundersen, L. L. Acta Chem. Scand. 1996, 50, 58–63.