Efficient Synthesis of an Adenosine A2a Agonist: Glycosylation of 2-Haloadenines and an N2-Alkyl-6-chloroguanine

Article abstract of DOI:10.1021/jo049963x

A convergent synthesis of adenosine A2a agonist 1 in the form of its maleate salt 2 was achieved. The key step in this approach was the highly selective 9β-glycosylation reaction between 2-haloadenines or an N ²-alkyl-6-chloroguanine and a D-ri

Full text of DOI:10.1021/jo049963x

Efficien t Syn th esis of a n Ad en osin e A2a
Agon ist: Glycosyla tion of 2-Ha loa d en in es
a n d a n N2-Alk yl-6-ch lor ogu a n in e
J ohn M. Caddell, Alan M. Chapman, Bob E. Cooley,
Brian P. Downey, Michael P. LeBlanc, Mary M. J ackson,
Thomas M. O’Connell, Hahn-My Phung,
Thomas D. Roper, and Shiping Xie*
Chemical Development, GlaxoSmithKline,
Research Triangle Park, North Carolina 27709
shiping.x.xie@gsk.com
Received J anuary 5, 2004
F IGURE 1. Retrosynthetic analysis.
Abstr a ct: A convergent synthesis of adenosine A2a agonist
1 in the form of its maleate salt 2 was achieved. The key
step in this approach was the highly selective 9â-glycosy-
lation reaction between 2-haloadenines or an N2-alkyl-6-
chloroguanine and a D-ribose derivative containing a 2-eth-
yltetrazolyl moiety. Glycosylations of other purine derivatives
were also examined, and the methods developed provide
efficient access to a variety of adenosine analogues such as
2-alkylaminoadenosines, an attractive class of compounds
with antiinflammatory activity.
Synthesis of 2-alkylaminoadenosines is generally ac-
complished through glycosylation of N2-alkylguanines4 or
2,6-dihalopurines5 followed by functionalization of the
resultant glycosides. Although there are a number of
options for carrying out selective glycosylation of purines,
the most attractive with regard to ease of operation and
potential for scalability appears to be the conditions
developed by Vorbruggen (the silyl Hilbert-J ohnson
reaction) which employ a silylated nucleoside base and
a glycosyl donor.6 Although often separable by chroma-
tography, significant amounts of undesired N7 glycosyl
and other isomers can be generated from glycosylation
of guanine by tetraacetylribose or other furanose deriva-
tives under standard Vorbruggen reaction conditions.7 In
contrast, glycosylation of 2,6-dihalopurines offers in-
creased selectivity of N9 â-anomer and opportunity to
access the desired 2-alkylaminoadenosines.
We now wish to report our studies directed toward the
synthesis of 2-alkylaminoadenosine 1 based on glycosy-
lation of several purine derivatives by 5-(2-ethyl-2H-
tetrazol-5-yl)tetrahydrofuran-2,3,4-triyl triacetate 3 as
shown in Figure 1. To our knowledge, there has been
little reported on the direct glycosylation of 2-halo-
adenines 4a and 4b,8,9 2-alkylaminoadenines such as 4c,
or N2-alkyl-6-chloroguanines such as 4d by ribofuranose
There is evidence from both in vitro and in vivo studies
to indicate that stimulation of adenosine A2a receptors
on immune cells produces a series of responses that can
be categorized as antiinflammatory.1,2 It is known that
adenosine and some analogues decrease neutrophil ad-
herence to biological substrates and diminish release of
tissue-destructive oxidative and nonoxidative products.2
There has been progressive development of compounds
that show increased potency and selectivity as an A2a
agonist through addition of 2-alkylamino substituents.3
Adenosine analogue 1 and its maleate salt 2, a 2-alkyl-
amino-9-(1-â-ribofuranosyl)adenine containing an eth-
yltetrazole moiety, are highly selective A2a agonists and
have been shown to be efficacious in several animal
models of inflammatory disease. Our interest in this class
of compounds, and the need for a large quantity of drug
substance to support both preclinical and clinical activi-
ties, prompted us to search for an efficient synthesis of
9â-2-alkylaminoadenosines such as 1.
(3) (a) Allen, D. G.; Chan, C.; Cousins, R. P. C.; Cox, B.; Geden; J .
V.; Hobbs, H.; Keeling, S. E.; Redgrave, A. J .; Roper, T. D. IV; Xie, S.
PCT Int. Appl., WO 9967265, 1999. (b) Cox, B.; Keeling, S. E.; Allen,
D. G.; Redgrave, A. J .; Barker, M. D.; Hobbs, H.; Roper, T. D., IV;
Geden; J . V. PCT Int. Appl., WO 9828319, 1998. (c) J arvis, M. F.;
Schulz, R.; Hutchison, A. J .; Do, U. H.; Sills, M. A.; Williams, M. J .
Pharmacol. Exp. Ther. 1989, 888.
(4) (a) Wright, G. E.; Dudycz, L. W. J . Med. Chem. 1984, 27, 175.
(b) Dudycz, L. W.; Wright, G. E. Nucleosides Nucleotides 1984, 3, 33.
(c) Robins, M. J .; Uznanski, B. Can. J . Chem. 1981, 59, 2601.
(5) (a) Marchand, A.; Lioux, T.; Mathe, C.; Imbach, J .-L.; Gosselin,
G. J . Chem. Soc., Perkin Trans. 1 1999, 2249. (b) Martin, P. Helv. Chim.
Acta 1996, 79, 1930.
(6) For reviews, see: (a) Vorbruggen, H. Acc. Chem. Res. 1995, 28,
509. (b) Vorbruggen, H. Acta Biochim. Pol. 1996, 43, 25.
(7) (a) Zou, R.; Robins, M. J . Can. J . Chem. 1987, 65, 1436. (b)
Robins, M. J .; Zou, R.; Hansske, F.; Madej, D.; Tyrell, D. L. J .
Nucleosides Nucleotides 1989, 8, 725.
(8) Synthesis of 2-haloadenosines has been reported through mi-
crobiological transglycosylation. For 2-fluoroadenine, see: (a) Na-
kayama, K.; Tanaka, H. J apan Patent 51027755, 1970; Chem. Abstr.
1985, 65, 2870. For 2-chloroadenine, see: (b) Mikhailopulo, I. A.;
Zinchenko, A. I.; Kazimierczuk, Z.; Barai, V. N.; Bokut, S. B.;
Kalinichenko, E. N. Nucleosides Nucleotides 1993, 12, 417. For
2-bromoadenine, see: (c) Huang, M.; Avery, T. L.; Blakley, R. L.;
Secrist, J . A., III; Montgomery, J . A. J . Med. Chem. 1984, 27, 800.
(1) For a recent review on medicinal chemistry of adenosine A2a
receptor agonists, see: Cristalli, G.; Lambertucci, C.; Taffi, S.; Vittori,
S.; Volpini, R. Curr. Top. Med. Chem. 2003, 3, 387.
(2) For a review on role of A2a adenosine receptors in inflammation,
see: Sullivan, G. W.; Linden, J . Drug. Dev. Res. 1998, 45, 103.
10.1021/jo049963x CCC: $27.50 © 2004 American Chemical Society
Published on Web 04/01/2004
3212
J . Org. Chem. 2004, 69, 3212-3215

SCHEME 1
SCHEME 2
derivatives.10 We were thus interested in investigating
the silyl J ohnson-Hilbert-type glycosylation of a selec-
tion of purines through which a more convergent syn-
thesis of 2-alkylaminoadenosines might be attained. Key
considerations for developing a scalable preparation
included the N9/N7 and â/R selectivity in the glycosylation
as well as the downstream efficiency in functionalization
of the glycosyl purines.
TABLE 1. Glycosyla tion of P u r in es 4a -f w ith
Tr ia ceta te 3a
Triacetate 3 was prepared from tetrazole 511 without
the need for chromatography as shown in Scheme 1.
Ethylation of tetrazole 5 with EtI and K2CO3 resulted in
a 2.4:1 ratio of the N2/N1 isomers (6/7).
The undesired N1 ethyltetrazolyl 7 was largely re-
moved as a solid by fractional crystallization of the crude
product from cyclohexane which provided a 92:8 N2/N1
mixture of tetrazoles 6 and 7 in 62% yield. The assign-
ment of N1 ethyl structure 7 was supported by an X-ray
crystallographic determination, and through the process
of elimination, the N2 ethyl structure for 6 was assigned.
Sequential cleavage of the acetonide, acetylation of the
resultant diol, and conversion of the resulting 2,3-di-O-
acetyl-1-O-methyl glycoside to the activated glycosyl
donor 3 were accomplished in 81% yield. Interestingly,
the N1 ethyltetrazolyl isomer of 3 was susceptible to
hydrolysis in the aqueous NaHCO3 workup that followed
the acid-catalyzed acetolysis. As a result, triacetate 3
containing less than 1% of N1 isomer was obtained in 81%
yield with a simple extractive workup. The two epimers
of 3 reacted with similar efficiency in the glycosylation
conditions when trimethylsilyl trifluoromethanesulfonate
(TMSOTf) was used for catalysis.
Substituted purine derivatives were prepared as shown
in Scheme 2. Reaction of 2-bromohypoxanthine12 with
L-phenylalaninol afforded guanine 4e in 74% yield.
Acetylation (68%) followed by chlorination with POCl3
(60%) led to 6-chloroguanine 4d after in situ hydrolysis
of the remaining acetyl protecting group. Finally, treat-
ment with ammonia provided 2,6-diaminopurine 4c in
85% yield.
9â
ratio
9â yieldc
purine
X
Y
R
product 9â/isomersb
(%)
4a
4b
4c
4d
4e
4f
F
Cl
NH2 Ac
NH2
9a
9b
9c
9d
9e
9f
29:1
23:1
9:1
93:1
1.8:1
14:1
76
83
78
71
47
78
H
Ph-AlaNH NH2 Ac
Ph-AlaNH Cl
Ph-AlaNH OH
Ac
Ac
Ac
Cl
Cl
a
All reactions were performed with 3 (1.2 equiv), TMSOTf, and
b
BSA in MeCN at reflux. Ratio was determined by HPLC-MS
analysis on either the reaction mixture or crude products. c 9a was
isolated by crystallization and 9b by crystallization as diol after
deacetylation in the presence of MeOH and K2CO3. All others were
isolated by chromatography.
silylation reagent and TMSOTf as a catalyst. Although
most literature procedures employ pre-silylation of the
purine base, we found that comparable results were
generally obtained with Wright’s protocol of direct mixing
of the peracetylated riboside and the purine base with
BSA and TMSOTf in refluxing MeCN.4a Results are
summarized in Table 1. Assignment of the 9â structures
as shown was based on subsequent conversion of the
major product to the target molecule 2, the structure of
which was confirmed by careful NMR studies including
NOE and HMBC (heteronuclear multiple bond correla-
tion) experiment. The ratio of major isomer to all minor
isomers was determined by HPLC-MS method. In this
method, all peaks in the HPLC chromatograms of the
quenched reaction mixture or the crude product were
identified by molecular weight. The integrals or areas of
all peaks of the same molecular weight as the desired
isomer were added together and considered the sum of
all minor isomers. The ratio was calculated by dividing
the area of the desired 9â isomer by the sum of all minor
isomers. This ratio determination assumes that all the
isomers have the same UV absorption efficiency at the
same detection wavelength due to structural similarity.
Coupling of triacetate 3 with purines 4a -f was carried
out with N,O-bis(trimethylsilyl)acetamide (BSA) as a
(9) An arabinofuranosyl 2-fluoroadenine was reported to be prepared
from a 1-chloroarabinose compound. See: Markovac, A.; Kalamas, R.
L.; Lamontagne, M. P. Patent AU/642611, 1993.
(10) Glycosylation of the Na or Hg salt of 6-chloroguanines by
1-chloro- and 1-bromoribofuranoses has been reported, mostly with low
yield and poor selectivity. See: (a) Hanna, N. B.; Ramasamy, K.;
Robins, R. K.; Revankar, G. R. J . Heterocycl. Chem. 1988, 25, 1899.
(b) Chu, C. K.; Matulic-Adamic, J .; Huang; J .-T.; Chou, T.-C.; Burch-
enal, J . H.; Fox, J . J .; Watanabe, K. A. Chem. Pharm. Bull. Med. 1989,
37, 336.
(11) Tetrazolyl compound 5 was prepared from a reaction of the
precursor nitrile with NaN3 in DMF. See: Schmidt, R. R.; Heermann,
D.; J ung, K.-H. Liebigs Ann. Chem. 1974, 1856.
(12) Beaman, A. G.; Gerster, J . F.; Robins, R. K. J . Org. Chem. 1962,
27, 986.
J . Org. Chem, Vol. 69, No. 9, 2004 3213

TABLE 2. Con ver sion of 9a -d to Ma lea te 2
2-haloadenosines was particularly practical owing to the
excellent crystallinity of the nucleoside products of the
glycosylation reaction. The chloro displacement of 2-chlo-
roadenosine 9b was somewhat difficult and required the
use of neat phenylalaninol as a melt. However, the
displacement with the more reactive 2-fluoro (9a ) and
6-chloro (9d ) nucleosides proved facile and provided
access to a variety of adenosine analogues with potential
for further functionalization in either the 2 or 6 positions
of the purine moiety. Separately, guanosine 9e and
ribofuranosyl 2,6-dichloropurine 9f were converted in
several steps to 2 following sequences previously reported
in the literature.4a,5a
substrate
X
Y
R
reagentsa and yield
9a
F
NH2 Ac (i) A; (ii) B; (iii) C.
75% for three steps
9b
Cl
NH2
H
(i) A; 83% from 4b; (ii) D; (iii) C.
In summary, we have achieved excellent 9â selectivity
in glycosylation of 2-haloadenines 4a and 4b, as well as
N2-alkyl-6-chloroguanine 4d by 5-(2-ethyl-2H-tetrazol-5-
yl)tetrahydrofuran-2,3,4-triyl triacetate 3. Moreover, these
glycosides have been converted to 2-alkylaminoadenosine
1 in the form of its maleate salt 2 by simple halogen
displacement at the 2- or 6-position of the purine moiety.
We believe that the sequences we described herein
provide efficient access to potentially a wide variety of
2-alkylaminoadenosines.
59% for two steps
Ph-AlaNH NH2 Ac (i) A; (ii) C. 62% for two steps
Ph-AlaNH Cl Ac (i) E; (ii) C. 83% for two steps
9c
9d
a
Key: (A) K2CO3, MeOH; (B) L-phenylalaninol, i-Pr2NEt,
DMSO, 100 °C; (C) maleic acid, MeOH/EtOH; (D) L-phenylalaninol
105 °C; (E) NH3, i-PrOH, 100 °C.
The ratio thus determined was in agreement with the
1H NMR of the crude products in cases where the ratio
was relatively low (4c, 4e and 4f). No efforts were made
to assign structures for the minor isomers which presum-
ably included N7 and R nucleosides.
Exp er im en ta l Section
Superior selectivity for the desired 9â isomer was
obtained with 2-fluoroadenine 4a 13 (29:1) and 2-chloro-
adenine 4b14 (23:1). The 93:1 selectivity with 6-chlo-
roguanine derivative 4d was particularly impressive.15
With the exception of 4e, all bases provided the desired
nucleosides in good isolated yield. In addition to being
formed with high stereoselectivity, the product 2-fluoro-
and 2-chloroadenosines were found to be highly crystal-
line, a property which significantly aided isolation and
purification in large scale. Although the use of diami-
nopurine 4c in the glycosylation reaction represents the
most convergent sequence to the desired compound, the
selectivity was comparatively low (9:1), as was the
selectivity with guanine 4e (1.8:1).
Vorbruggen has postulated that the initial kinetically
formed N3 isomer rearranges to the thermodynamic N9
nucleoside via the intermediacy of the N7 nucleoside.6 The
ratio of N9 to N7 products obtained is presumably related
to the relative thermodynamic stability of the two
isomeric nucleosides. There has been a mechanistic study
on isomer distribution in the formation of 6-oxopurines
such as 4e.4a In view of the excellent results obtained
with use of halogenated purines 4a , 4b, and 4d , studies
are ongoing to determine the basis for the observed
selectivity among various purine derivatives.
Repr esen tative P r ocedu r e for Glycosylation : (2R,3R,4R,-
5R )-2-(6-Am in o -2-flu o r o -9H -p u r in -9-y l)-5-(2-e t h y l-2H -
tetr a zol-5-yl)tetr a h yd r ofu r a n -3,4-d iyl Dia ceta te (9a ). To a
mixture of 65.0 g (190 mmol) of triacetate 3 and 24.1 g (157
mmol) of 2-fluoroadenine 4a in 372 mL of MeCN was succes-
sively added 58.5 mL (238 mmol) of N,O-bis(trimethylsilyl)-
acetamide (BSA) and 34.5 mL (190 mmol) of TMSOTf at ambient
temperature. The yellow suspension was heated at reflux and
for 5 h. After being cooled to ambient temperature, the reaction
was quenched with 390 mL of 10% KHCO3 and extracted with
CH2Cl2 (3 × 300 mL). The combined organic layers were washed
with 400 mL of 10% brine, filtered through a short pad of Celite,
and partially concentrated in vacuo to 300 mL. The resultant
white slurry was treated with 500 mL of EtOH and concentrated
to 350 mL. The product was filtered off and dried in vacuo at 50
°C to afford 51.9 g (76%) of 2-fluroadenosine 9a as a white
crystalline solid. Before the reaction was quenched, an aliquot
was taken and diluted with H2O-MeCN (1:9). This crude
mixture was analyzed by HPLC-MS. The analysis indicated
90% of desired product 9a along with 3.1% total of five isomeric
impurities. The ratio of the desired 9 â product and isomers was
thus 29:1. 9a : mp 208-210 °C dec; [R]20 -35 (c 1.1, MeOH);
D
IR 1763, 1673, 1610 cm-1
;
1H NMR (CDCl3) δ 1.69 (t, J ) 7.5
Hz, 3H), 2.07 (s, 3H), 2.21 (s, 3H), 4.74 (q, J ) 7.5 Hz, 2H), 5.58
(s, 1H), 5.78 (dd, J ) 4.8, 2.4 Hz, 1H), 5.93 (br s, 2H), 6.16 (dd,
J ) 6.7, 4.8 Hz, 1H), 6.46 (d, J ) 6.7 Hz, 1H), 8.39 (s, 1H). Anal.
Calcd for C16H18FN9O5: C, 44.14; H, 4.17; N, 28.95. Found: C
44.09; H, 4.13; N, 28.78.
Repr esen tative P r ocedu r e for P r epar ation of 2: (2R,3R,-
4S,5R)-2-(6-Am in o-2-[[(1S)-2-h yd r oxy-1-(p h en ylm et h yl)-
et h yl]a m in o]-9H -p u r in -9-yl)-5-(2-et h yl-2H -t et r a zol-5-yl)-
tetr a h yd r o-3,4-fu r a n d iol Ma lea te Sa lt (2). A mixture of 23.1
g (53.1 mmol) of 2-fluoroadenosine 9a and 14.7 g (106 mmol) of
K2CO3 in 215 mL of MeOH was stirred at ambient temperature
overnight. The resultant slurry was diluted with 300 mL of
EtOAc and poured into 250 mL of water. The aqueous layer was
extracted with EtOAc (4 × 200 mL). The combined organic layers
were dried with anhydrous Na2SO4, concentrated in vacuo, and
dried at 60 °C to give 18.5 g (99%) of the 2,3-dihydroxy derivative
of 2-fluoroadenosine 9a as a white solid. A mixture of 18.2 g
(52.0 mmol) of this solid and 16.5 g (109 mmol) L-phenylalaninol
in 40 mL of DMSO and 120 mL of i-Pr2NEt was heated at 100
°C overnight. The mixture was cooled to 40 °C, diluted with 200
mL of water, and extracted with 20:1 EtOAc-MeOH (4 × 200
Conversion of 9a -d to 1 was accomplished as shown
in Table 2, and the product was isolated as the crystalline
maleate salt 2 by simple filtration of the reaction mixture.
The assignment of relative structure of 2 was confirmed
by NMR analysis which included studies including NOE
and HMBC (heteronuclear multiple bond correlation)
experiment. The absolute stereochemistry was assigned
based on the starting D-ribose structure. Processing of
(13) Fluoroadenine was purchased from either Allied Signal Spe-
cialty Chemicals or Fluorochem Ltd.
(14) Brown, G. B.; Weliky, V. S. J . Org. Chem. 1958, 23, 125.
(15) Similar selectivity was observed when 4a and 4b were
glycosylated by â-D-ribofuranose 1,2,3,5-tetraacetate.
3214 J . Org. Chem., Vol. 69, No. 9, 2004

mL). The combined organic layers were washed with 200 mL of
brine and concentrated in vacuo. The resultant crude product
(39 g) was chromatographed on silica gel. Elution with 10%
MeOH in CH2Cl2 afforded 21.2 g (85%) of 1 as a light yellow
solid. A solution of 1.29 g (2.68 mmol) of free base 1 and 311 mg
(2.68 mmol) of maleic acid in 16 mL of 95:5 EtOH-MeOH was
stirred at ambient temperature for 4 h. The resultant white solid
was filtered, washed with 3 mL of EtOH, and dried in vacuo at
60 °C to provide 1.42 g (89%) of maleate salt 2 as a white
7.34 (m, 5H), 8.27 (s, 1H). Anal. Calcd for C25H30N10O8: C,
50.16: H, 5.07; N, 23.40. Found: C 50.04; H, 5.07; N, 23.33.
Ack n ow led gm en t. We are grateful to Dr. Sean
Lynn for performing the X-ray crystallographic analysis.
We thank Dr. J ohn Roberts for helpful discussions
during preparation of the manuscript.
Su p p or tin g In for m a tion Ava ila ble: Detailed experi-
mental procedures and analytical information for compounds
2, 3, 4c-e, 6, and 9a -f, X-ray crystallographic data for 7, and
detailed NMR analysis for 2. This material is available free
of charge via the Internet at http://pubs.acs.org.
crystalline solid: mp 169 °C dec; [R]25 -30.0 (c 1.0, MeOH); IR
D
1705, 1663, 1628, 1508 cm-1; 1H NMR (MeOH-d4) δ 1.63 (t, J )
7.3 Hz, 3H), 2.97 (m, 2H), 3.33 (m, 1H), 3.68 (m, 2H),4.38 (m,
1H), 4.68 (m, 1H), 4.72 (q, J ) 7.3 Hz, 2H), 4.83 (m, 1H), 5.36
(d, J ) 4.5 Hz, 1H), 6.12 (d, J ) 5.8 Hz, 1H), 6.30 (s, 2H), 7.13-
J O049963X
J . Org. Chem, Vol. 69, No. 9, 2004 3215